top incomes in the long run of history - thomas...
TRANSCRIPT
13
Top incomes in the long run of history
A B Atkinson, Thomas Piketty and Emmanuel Saez
13.1 INTRODUCTION
In this book, and Volume 1 (Atkinson and Piketty, 2007), we have assembled
evidence on top incomes in 22 countries, covering periods that range from
15 years (China) and 30 years (Italy) to 120 years (Japan) and 132 years
(Norway). The coverage by countries and years is shown in Figure 13.1. For
7 of the 22 countries, the series start before the First World War (1914), and
for all but 3 we have observations from before the Second World War. The
median number of observations per country is 67.5. To avoid any possible
misunderstanding, we should make clear that is not a rectangular data set:
there are many missing years (if any reader can help fill the gaps, we should
be pleased to hear from them). In fact, the total number of country/year
observations is 1,454.
By the standards of economics, these are a rich set of time series.
The results for top income shares for the 22 countries are summarised in the
Appendix Tables 13A.1 to 13A.22. Estimates of the Pareto-Lorenz
coefficients (and the inverse coefficients) characterizing the upper tail of
2
the distribution in these 22 countries are provided in Appendix Tables
13A.23 and 13A.24. In the case of the 10 countries covered in the first
volume, we have been able to extend the series in some cases. We have also
incorporated (under Germany) the results for Prussia for the period 1891 to
1919 (see Dell, 2008). The estimates reported in these Appendix tables are
those excluding capital gains (where these can be excluded), an aspect
discussed below in Section 13.2, and they relate to gross incomes, typically
after transfers but before the deduction of income tax. It should be noted
that we focus in this chapter on the studies of top incomes published in
these two volumes, but there have been other important recent
contributions, including those by Merz, Hirschel and Zwick (2005) and by
Bach, Corneo and Steiner (2008) of Germany, by Gustafsson and Jansson
(2007) of Sweden, and by Guilera (2008) of Portugal. The reader is also
referred to the valuable survey by Leigh (2009).
In this chapter, we summarise the main findings from the two
volumes, and consider the range of possible explanations. We first, in
Section 13.2, discuss the quality and comparability of the data. How far can
the results from different chapters be treated as comparable? What are the
limitations of income tax data as a source? The qualifications set out in this
section are essential reading before making use of the results on top income
shares, presented in summary form in Section 13.3. Our summary starts
with the sixty years since the Second World War, taking 1949 as a point of
departure, before turning to the lessons from the years before 1949. We
ask how far there are common patterns of change. We ask whether top
incomes are different. In the final part of the chapter (Sections 13.4 and
3
13.5), we turn to possible explanations for the changes in top shares over
time. In Section 13.4, we consider how the behaviour and functional form of
top income shares can be modelled – theoretically and empirically. In
Section 13.5, we examine some of the main forces that have been evoked in
the individual chapter histories in both this and the previous volume. These
include the impact of progressive taxation, globalisation, and political
change. The final section 13.6 concludes.
In considering the 22 countries, it is natural to look for groupings. In
the first volume, we contrasted the English-speaking countries and
Continental Europe. In the present volume, we have added three Nordic
countries. Are they part-way between being Anglo-Saxon and Continental?
We have added three Southern European countries. Do they have common
features? We have added five Asian countries, two with high GDP per
capita, and three less developed but rapidly growing. Are top income shares
rising with fast economic growth? Or does the behaviour of their top shares
reflect a global pattern? These questions are discussed in more detail
below, but in Table 13.0 we summarise in words the main findings from the
country chapters in this volume.
4
Table 13.0 Summary of main findings from Chapter 1 to 12. Country Main findings 1. India The shares of the top 0.01%, the top 0.1% and the top 1% shrank
substantially from the 1950s until the early to mid 1980s but then went back up again, so that today these shares are only slightly below the level of the 1920s-30s. This U-shaped pattern is broadly consistent with the evolution of economic policy in India: the period from the 1950s to the early to mid 1980s was also the period of “socialist” policies in India, while the subsequent period saw a gradual shift towards more pro-business policies.
2. China Income inequality in urban China has increased at a high rate between 1986 and 2003, with the share of the top 10 per cent increasing by more than 60 per cent. The share of the top 1 per cent more than doubled.
3. Japan Income concentration was extremely high throughout the pre-Second World War period during which the nation underwent rapid industrialization; a drastic de-concentration of income at the top took place in 1938-1945; income concentration remained low during the rest of the century but shows some sign of increase in the last decade; and top income composition in Japan has shifted dramatically from capital income to employment income over the course of the 20th century.
4. Indonesia
Top income shares grew during the 1920s and 1930s, but fell in the post-war era. In more recent decades, there was a sharp rise in top income shares during the late 1990s, coinciding with the 1997-98 economic crisis, and some evidence that top income shares fell in the early-2000s.
5. Singapore
During the time Singapore was a colony, top income shares rose to a peak in 1951 and then declined over the 1950s. Following independence there followed 25 years of broad stability at the very top. The 1990s saw a fall in top shares, but after the Asian financial crisis they rose by around a half, and they remain above earlier levels.
6. Argentina
There was an increase in top income shares after the Great Depression, with maxima in 1942-1944, and a substantial decline during the Peronist years. However, the limits of the Peronist redistributive policy are marked by the fact that in 1956, if lower than in 1945, the top shares were still above the ones observed in the developed world; they were higher than in the United States, France, Australia and even Spain. Since the mid 1990s, top income shares followed an increasing trend, similar to the pattern found in Anglo-Saxon economies.
7. Sweden Starting from levels of inequality approximately equal to those in other Western countries at the time, the income share of the Swedish top decile dropped sharply over the first 80 years of the twentieth century. Most of the decrease takes place before the expansion of the welfare state and by 1950 Swedish top
5
income shares were already lower than in other countries. The fall is almost entirely due to a dramatic drop in the top percentile explained mostly by decreases in capital income, while the lower half of the top decile – consisting mainly of wage earners – experienced virtually no change over this period. In the past decades top income shares evolve very differently depending on whether capital gains are included or not. With capital gains included, Sweden’s experience resembles that in the U.S., with sharp increases in top incomes; excluding capital gains, Sweden looks more like the continental European countries.
8. Finland The share of the top 1 per cent of income earners declined from around 15 per cent of taxed income among the population of tax units in 1920 to just over 6 per cent in the late 1940s to rapidly increase to about 10 per cent in the later 1950s. The top 1 per cent share then declines for almost thirty years until the early 1990s. The increase in this top share in the late 1990s is steep and brings it in 2000, when it peaked, to the same level as seen in the 1950s. The total share of the highest earners fell consistently from the beginning of the 1960s to the mid 1990s but then began to rise. The results bring out clearly how the major equalization from the beginning of 1960 to the mid 1990s has been reversed, taking the shares of top income groups back to levels of inequality or even higher than those found over 40 years ago.
9. Norway Top income shares in the nineteenth century were high: the share of the top 1 per cent was around 20 per cent and that of the top 0.5 per cent around 15 per cent. There was a rise in the shares of the top 10 per cent, 5 per cent and 1 per cent between 1875 and 1888. Between 1896 and 1902 there was a definite fall; there was some recovery in 1906, but then a further fall. After (and during) the First World War, there was some recovery in the top shares. There was a sharp drop between the 1930s and the late 1940s. Over the early part of the post Second World War period, top income shares fell steadily: the share of the richest 0.5 per cent fell from 6.4 per cent in 1948 to 2.8 per cent in 1991. There was then, as in Sweden, the UK and the US, a turning point, which came at the start of the 1990s, rather later than in the UK and the US.
10. Spain Income concentration was much higher during the 1930s than it is today. The top 0.01% income share was twice higher in the 1930s than in recent decades. The top 0.01% income share fell sharply during the first decade of the Franco dictatorship, and has increased slightly since the 1970s, and especially since the mid-1990s. Both the level and the time pattern of the top 0.01% income share in Spain are fairly close to comparable estimates for the US and France over the period 1933-1971, especially the decades after the Second World War.
11. Income concentration was, as in Spain, much higher during the
6
Portugal 1930s and 1940s than it is today. Top income shares stayed relatively stable between the end of the Second World War and the end of the 1960s, followed by a large drop. The drop began to be reversed at the beginning of the 1980s. Over the last fifteen years top income shares have increased steadily, with a considerable contribution from the rise in wage dispersion.
12. Italy There has been a persistent increasing pattern in top income shares since the mid 1980s, mainly driven by top wages and self-employment income. Notwithstanding the increasing trend, the rise in Italian top shares has been small relative to the surge experienced by top incomes in the United States and other Anglo-Saxon developed economies
7
13.2 QUALITY AND COMPARABILITY OF THE DATA
The evidence presented in this, and the preceding, volume comes almost
exclusively from tax and other administrative returns in different countries.
The use of tax data is often regarded by economists with considerable
disbelief. The index to Morgenstern’s book On the Accuracy of Economic
Observations (1963) contains the entry “income tax, as reason for lying”,
and this summarizes well the general – if not very specific - skepticism. In
the UK, Richard Titmuss wrote around the same date a book-length critique
of the income tax-based statistics on distribution, concluding, “we are
expecting too much from the crumbs that fall from the conventional tables”
(1962, page 191). More recently, compilers of databases on income
inequality have tended to rely on household survey data, dismissing income
tax data as unrepresentative.
These doubts are well justified for at least two reasons. The first is
that tax data are collected as part of an administrative process, which is not
tailored to our needs, so that the definition of income, of income unit, etc
are not necessarily those that we would have chosen. This causes particular
difficulties for comparisons across countries, but also for time-series
analysis where there have been substantial changes in the tax system, such
as the moves to and from the joint taxation of couples. Secondly, it is
obvious that those paying tax have a financial incentive to present their
affairs in such a way that reduces tax liabilities. There is tax avoidance and
tax evasion. The rich, in particular, have a strong incentive to understate
their taxable incomes. Those with wealth take steps to ensure that the
8
return comes in the form of asset appreciation, typically taxed at lower
rates or not at all. Those with high salaries seek to ensure that part of their
remuneration comes in forms, such as fringe benefits or stock options, that
receive favourable tax treatment. Both groups may make use of tax havens
that allow income to be moved beyond the reach of the national tax net.
These shortcomings limit what can be said from tax data, but this
does not mean that the data are worthless. Like all economic data they
measure with error the “true” variable in which we are interested. As with
all data, there are potential sources of bias, but, as in other cases, we can
say something about the possible direction and magnitude of the bias.
Moreover, we can compensate for some of the shortcomings of the income tax
data. It is true that income tax data cover only the taxpaying population,
which, in the early years of income tax, was typically only a small fraction of
the total population. However, following the pioneering contribution of
Kuznets (1953), we can combine the tax data with external estimates of the
total population and the total income. These control totals are typically based
on Censuses of Population and on national accounts estimates of the total
income of persons. The control totals require a number of adjustments and
are surrounded by a margin of error, but the important point is that when we
refer to the top 1 per cent having x per cent of income, this means the top 1
per cent of the total population and x per cent of the total income. It is not
the top 1 per cent of taxpayers. Nor is the total of income limited to that
accruing to taxpayers. We may not be able to describe the whole distribution,
but we can estimate the upper part of the Lorenz curve.
9
But why not use household surveys, which cover the whole (non-
institutional) population? Why use income tax data? There are two main
answers. The first is that household surveys themselves are not without
shortcomings. These include sampling error, which may be sizeable with the
typical sample sizes for surveys, whereas tax data drawn from administrative
records are based on very much larger samples. Indeed, in many cases the tax
statistics relate to the whole universe of taxpayers, in which case traditional
confidence intervals do not apply. Household surveys suffer from differential
non-response and incomplete response (these two being the survey
counterpart of tax evasion). Such problems particularly affect the top income
ranges, as is recognised in studies that combine household survey data with
information on upper income ranges from tax sources (see, for example, in
the UK, Brewer, Muriel, Phillips and Sibieta, 2008). The second answer is that
household surveys are a fairly recent innovation. Household surveys only
became regular on most countries in the 1970s or later, and in a number of
cases they are held at intervals rather than annually. We are interested here
in covering the whole post war period, and indeed in going back further. This
is what the income tax data allow us to do. The beauty of income tax
evidence is that it is available for long runs of years, typically on an annual
basis, and that it is available for wide variety of countries.
Comparability of methods
Although the authors of individual chapters in this volume and in Volume 1
have modelled their research on Piketty (2001), they have in some cases
been unable to follow exactly the same methods and in other cases they
10
have chosen a different approach. Some of these differences in
methodology are unlikely to affect the broad conclusions drawn, as has been
shown by sensitivity analysis in individual chapters. This applies to the
choice of interpolation method, which, at least within intervals (as opposed
to extrapolation of an open interval), is not going to have a major impact
(see Appendix 9C in Volume 1). The same applies to the choice of age cut-
off for the adult population. The studies for Australia, Finland, New
Zealand, Singapore, and the UK use persons aged 15 and over, while those
for Argentina, Canada, Italy, Japan, Portugal and the US use persons aged
20 and over, which means that the former may give a higher estimate of the
share of the top x per cent (since they are including more people from the
tax returns). However, the effect is small: Atkinson and Leigh (Chapter 7 in
Volume 1) find for Australia that using persons aged 20 and over would
reduce the share of the top 1 per cent by approximately a twentieth.
Other differences are quantitatively more important. Three of the
differences seem to us to be of particular significance. The first is the
difference in the unit of analysis. For Argentina, Australia, Canada, China,
and Italy (in the period covered), the unit is the individual. In a number of
other countries, including France, Germany, Ireland, the Netherlands,
Portugal, Switzerland and the US, the unit of analysis is the “tax unit”
combining the incomes of husbands and wives. In India, the unit is the
individual or the Hindu undivided family. In the case of China and Indonesia,
the estimates relate to households. The differences between these units of
analysis affect the comparability of the estimates in a way that depends on
the joint distribution of income. The difference could go in either direction
11
(see Atkinson, Chapter 2 in Volume 1). The unit may also change over time,
as in Japan (with adjustments made in Chapter 3 to the pre-1950 data), New
Zealand (since 1953), Spain (with adjustments made in Chapter 10) and the
UK (since 1990).
The differences in unit cannot be treated simply as a fixed effect.
The growth of female labour force participation means that the joint
distribution of earned incomes is now of much greater significance. The
ageing of the population means that there are more single elderly persons in
the distribution. On the other hand, we can learn from the cases where
there was a change. In the case of the US, Piketty and Saez (Chapter 5 in
Volume 1) increase the recorded income shares by “about 2.5 per cent” for
the earlier period 1913-1947 when there was a degree of separate filing
(Piketty and Saez, 2001, 35n).1 In the case of the United Kingdom, the
introduction of independent taxation in 1990 was associated with (Chapter 4
in Volume 1) a rise in the share of the top 1 per cent of around an eighth. In
the case of New Zealand (Chapter 8 in Volume 1), the introduction of
individual taxation in 1953 was associated with an upward jump of around a
quarter in the share of the top 1 per cent. Not all of this change can
necessarily be attributed to the introduction of independent taxation, but it
suggests that the difference between individual and tax unit bases needs to
be taken into account in interpreting the series for the different countries.
The second significant difference is in the derivation of control totals
for income. As described in Chapter 2, there are two main approaches. 1 It should be noted that they use throughout a control total based on tax units, so that
separate filing will definitely cause the top share to be understated.
12
These are illustrated by those applied in the US at different dates. Piketty
and Saez (Chapter 5 in Volume 1) for the second half of the period (from 1944)
extrapolate from the recorded incomes, imputing to non-filers a fixed
fraction of filers’ average income, to arrive at a total (tax defined) income
for all individuals. They note that the resulting total series is a broadly
constant percentage (between 77 per cent and 83 per cent) of total
personal income recorded in the national accounts if transfers are excluded.
They therefore take for the earlier period 1913-1943 a control total equal to
a constant percentage (80%) of total personal income less transfers. (The
estimates for Switzerland involve a similar combination of the two
approaches.) These two methods – estimates of the income of non-filers,
and national accounts-based totals – are used to differing degrees in
different countries. In the UK (Chapter 4 in Volume 1), the total income of
non-filers is constructed from estimates of the different elements of income
missing from the tax returns. The resulting total declines from around 95%
to around 85% of total personal income minus transfers recorded in the
national accounts. In the Netherlands, a similar approach is followed, with
similar implications for the relationship between the control total and total
personal income in the national accounts. In their estimates for Sweden
(Chapter 7), Roine and Waldenström compare the two methods. They make
estimates of total personal income from the categories in the national
accounts (the total varies around 70 per cent of GDP); they make estimates
by adding to the total in tax returns amounts for income not included and
estimates of the income of non-filers. They argue that the latter gives too
high a figure in the early years of the twentieth century, but that from 1930
13
it is close to 89 per cent of the personal income total. They therefore use
throughout a figure of 89 per cent of personal income, which is around 63
per cent of GDP. The approach followed in Norway (Chapter 9) is similar,
with a lower percentage (72 per cent) to correspond to the concept of
assessed income in the tax data.
Most of the estimates are based on the second approach, applied in
varying degrees of detail. In Canada, Saez and Veall (Chapter 6 in Volume
1) use throughout (1920 to 2000) a constant percentage (80 per cent),
applied to “total personal income less transfers”. The estimates for Ireland
(Chapter 12 in Volume 1) follow the same method. For Japan (Chapter 3),
Moriguchi and Saez construct a personal income total from the national
accounts, deducting items that do not appear in taxable income such as
employer social contributions and imputed rents on owner-occupied homes.
For Spain (Chapter 10), Alvaredo and Saez add for 1981-2004 the national
accounts figures for wages and salaries (not including social contributions),
plus 50 per cent of transfers, plus two-thirds of unincorporated business
income, plus all non-labour non-business income reported on tax returns.
This yields a figure around 66 per cent of GDP and they apply this
percentage in the earlier years 1933-1971. For Portugal (Chapter 11) in
1989-2003 and Italy (Chapter 12) the procedure is similar. For Portugal, the
percentage of GDP is 60 per cent, and this is assumed to apply to the earlier
period 1936-1983.
In considering the control totals for income, we need to bear in mind
that the present volume covers countries where national accounting has
been more recently developed and where historical data are hard to obtain.
14
As we saw in Chapter 5, the national income figures for the 1940s and 1950s
in Singapore involved “a considerable amount of guesswork”. This has meant
adopting more approximate methods. The control totals for India (Chapter
1) are taken as 70 per cent of national income. In the case of Singapore,
allowance has to be made for the international position of the economy,
and the estimates are based on figures for indigenous GDP, taking a variable
percentage to represent household income. The Indonesian estimates of
Leigh and van der Eng (Chapter 4) make use of input output data.
The differences in method are greatest in the area of income totals,
and the resulting estimates of top income shares need to be treated with
caution. If, for example, the appropriate income total were considered to
be 60 per cent rather than 70 per cent of GDP, the top share would be
increased by more than 15 per cent. At the same time, we should note that
the estimates of shares-within-shares, and the (inverse) Pareto-Lorenz
coefficients, are not affected by differences in the income totals, since they
are measures of the shape of the upper part of the distribution.
The definition of taxable income
Taxes affect the substance of the income distribution, and we return to this
in Section 13.3, but they also affect the form of the income distribution
statistics. In all cases, the estimates follow the tax law, rather than a
“preferred” definition of income. The latter income concept may seek to
approximate the Haig-Simons comprehensive definition, including such
items as imputed rent, in kind employment benefits, capital gains and
15
losses, and transfer payments.2 For a single country study, it may be
reasonable to assume that taxable income is a concept well understood in
that context. Alternatively, one may assume that all taxable incomes differ
from the preferred definition by the same percentage. Neither of these
assumptions, however, seems particularly satisfactory, and use of taxable
income may well affect the conclusions drawn about changes over time.
When we come to a cross-country comparison, there seems an even
stronger case for adopting a definition of income that is common across
countries and that does not depend on the specificities of the tax law in
each country.
Approaching a common definition of income does however pose
considerable problems, as illustrated by the treatment of transfers (which
have grown very considerably in importance over the century), by capital
gains, by the inter-relation with the corporate tax system, and by tax
deductions. The studies for the US and Canada subtract social security
transfers on the grounds that they are either partially or totally exempt
from tax. In other countries, such as Australia, New Zealand, Norway, and
the UK, the tax treatment of transfers differs, with typically more transfers
being brought into taxation over time.
Perhaps the most important aspect that affects the comparability of
series overtime within each country has been the erosion of capital income
from the progressive income tax base. Early progressive income tax systems 2 In principle, transfers from the government should be not be included in pre-fisc incomes as they are part of the government redistributive schemes which tax pre-fisc incomes and provide transfers. In practice, the largest cash transfer payments are public pensions which are often related to social security contributions during the work life and hence can be considered as deferred earnings. Means-tested transfer programs are in general non-taxable and excluded from the estimates presented.
16
included a much larger fraction of capital income than the present
progressive income tax systems. Indeed, overtime, many sources of capital
income, such as interest income, or returns on pension funds, have been
either taxed separately at flat rates or fully exempted, and hence have
disappeared from the tax base. Some early income tax systems (such as
France from 1914 to 1964) also included imputed rents of home owners in
the tax base, but today imputed rents are typically excluded. As a result of
this imputed rent exclusion and the development of numerous other forms
of legally tax-exempt capital income, the share of capital income that is
reportable on income tax returns, and hence included in the series
presented, has significantly decreased overtime. To the extent that such
excluded capital income accrues disproportionately to top income groups,
this will lead to an underestimation of top income shares. Ideally, one
would want to impute excluded capital income back to each income group.
Because of lack of data, such an imputation is very difficult to fully carry
out. Some of the studies discuss whether the exclusion of capital income
affects the series. For example, Moriguchi and Saez in the case of Japan,
use survey data to try and assess to estimate how interest income - today
almost completely excluded from the comprehensive income tax base — is
distributed across income groups. In the case of France, Piketty has shown
that the long –run decline of top income shares was robust, in the sense that
even an upper bound imputation of today’s tax-exempt capital incomes to
today’s reported top incomes would be largely insufficient to undo the
observed fall. We should make clear however that there was no systematic
attempt to impute full capital income on a comparable basis over time and
17
across countries. We view this as one of the main shortcomings – probably
the main shortcoming – of our data set. As we shall see in Sections 13.4 and
13.5 below, this puts strong limitations on the extent to which one can use
our data set to rigorously test the theoretical economic mechanisms at play.
The treatment of capital gains and losses also differs across time and
across countries. In the US, “the tax treatment of capital gains and losses
has undergone several sweeping revisions since 1913” (Goode, 1964, page
184). Capital gains have been regarded as within the purview of the income
tax, but with different treatments regarding the deductibility of losses and
the rates of taxation. In Volume 1, Chapters 5 and 6 present series for the
US and Canada both excluding and including realised capital gains, and the
same procedure has been followed here for Japan (Chapter 3), Sweden
(Chapter 7), Finland (Chapter 8), and Spain (Chapter 10). The effects of the
inclusion of capital gains on the share of the top 1 per cent in the period
since 1949 are shown in Figure 13.2 for 5 of these countries (data are given
in Chapter 3 for Japan but only for the top 0.1 per cent). The adjustments
have been important in the US throughout the period but have increased in
recent years. In 1949, the exclusion of capital gains reduced the share of
the top 1 per cent by about a tenth; fifty years later, in 1999, it reduced
the share by about a fifth. In the case of Sweden, Roine and Waldenström
note that “over the past two decades the general picture turns out to
depend crucially on how income from capital gains is treated. If we include
capital gains, Swedish income inequality has increased quite substantially;
when excluding them, top income shares have increased much less”
(Chapter 7). The estimates for Spain (Chapter 10) show the share of the top
18
1 per cent rising between 1982 and 2002 from 6.9 to 8.5 per cent before
capital gains but from 7.0 to 9.5 per cent when capital gains are included.
Income tax systems differ in the extent of their provisions allowing
the deduction of such items as interest paid, depreciation, pension
contributions, alimony payments, and charitable contributions. Income from
which these deductions have been subtracted is often referred to as “net
income”. (We are not referring here to personal exemptions.) The aim is in
general to measure gross income before deductions, but this is not always
possible. The French estimates (Chapter 3 in Volume 1) show income after
deducting employee social security contributions. In a number of countries,
the earlier income tax distributions refer to income after these deductions,
but the later distributions refer to gross income. In the US, the income tax
returns prior to 1944 showed the distribution by net income, after
deductions. Piketty and Saez (Chapter 5 in Volume 1) apply adjustment
factors to the threshold levels and mean incomes for the years 1913-1943
(see Piketty and Saez, 2001, page 40). In Canada, the tax returns for 1920
to 1945 relate to net income. Deductions were smaller, and Saez and Veall
(Chapter 6 in Volume 1) make no adjustment prior to 1929 and for 1929 to
1945 increase all amounts by 2%. In Australia (Chapter 7 in Volume 1),
estimates for 1921-44 are based on taxable rather than total income by
ranges of taxable income, while the estimates from 1947-57 are based on
the distribution of taxable income by ranges of actual income. Using
estimates from overlapping years, adjustments are made to account for
these changes. The estimates for Norway in Chapter 9 relate to “assessed”
income after deduction of interest paid and certain other deductions.
19
The areas highlighted above – transfers, tax-exempt capital income,
capital gains, and deductions – may all give rise to cross-country differences
and to lack of comparability over time in the income tax data. Any user
needs to take them into account. The same applies to tax evasion, to which
we devote the next sub-section.
Tax avoidance and tax evasion
As highlighted by the quotation from Morgenstern, the standard objection to
use of income tax data to study the distribution of income is that tax
returns are largely works of fiction, as taxpayers seek to avoid and evade
being taxed. The under-reporting of income can affect cross-country
comparisons where there are differences in prevalence of evasion and can
affect measurement of trends where the extent of evasion has changed over
time.
It is not a coincidence that the development of income taxation
follows a very similar path across the countries studied in these volumes. All
countries start with progressive taxes on comprehensive income using high
exemption levels which limits the tax to only a small group at the top of the
distribution. Indeed, at an early stage of industrial development, when a
substantial fraction of economic activity takes place in small informal
businesses, it is just not possible for the government to enforce a
comprehensive income tax on a wide share of the population.3 However,
even in early stages of economic development, Alvaredo and Saez note “the
incomes of high income individuals are identifiable because they derive 3 Even today in the most advanced economies, small informal businesses can escape the individual income taxes.
20
their incomes from large and modern businesses or financial institutions
with verifiable accounts, or from highly paid (and verifiable) salaried
positions, or property income from publicly known assets (such as large land
estates with regular rental income).”4 Comprehensive income taxes are
extended to larger groups only when economic development has reduced
the number of untaxable informal income earners to a reasonably small
fraction of the population. Therefore, it is conceivable that the early
progressive income taxes, upon which statistics those studies are based,
captured reasonably well most components of top incomes.
The extent of contemporary tax evasion is considered specifically in a
number of the country chapters. In the case of Sweden, Roine and
Waldenström (Chapter 7) conclude that overall evasion is modest (around 5
percent of all incomes) and that there is no reason to believe that under-
reporting has changed dramatically over time. A speculative reason for this
may be that while the incentives to underreport have increased as tax rates
have gone up over time the administrative control over tax compliance has
also been improved. The Nordic countries may well be different. In the case
of Spain (Chapter 10), Alvaredo and Saez note the widely held view that
income tax evasion in Spain is (and was) very high, and that consequently,
the income tax data vastly under-estimate actual incomes. They go on to
examine the evidence for this view. Of course, such evidence is hard to
come by, and may only be partial, but it does exist. For instance, the
Spanish tax administration made public the list of taxpayers for tax years 4 Indeed, before comprehensive taxation starts, most countries had already adopted schedular separate taxes on specific income sources such as wages and salaries, profits from large businesses, rental income from large estates. Such taxes emerge when economic development makes enforcement feasible.
21
1933 to 1935, and from this it can be seen that virtually all the largest
aristocratic real estate owners were taxpayers. More generally, a careful
analysis of the income tax statistics shows that evasion and avoidance in
Spain at the very top of the distribution during the first decades of
existence of the tax was most likely not significantly higher than it was in
other countries such as the United States or France. In the case of Italy,
Alvaredo and Pisano note the widespread view of tax evasion being much
higher than in other OECD countries. Audits and subsequent scandals
involving show-business people, well-known fashion designers and sport
stars help support this idea among the general public, even when they also
provide evidence about the fact that top income earners are very visible for
the tax administration. The evidence for Italy does indeed suggest that
evasion is important among small businesses and the self-employed
(traditionally numerous in Italy), for whom there is no double reporting, but
that for wages, salaries and pensions at the top of the distribution there is
little room for evading those income components that must be reported
independently by employers or the paying authorities. They conclude that
the evasion from self-employment and small business income is unlikely to
account for the gap in top incomes between Italy and Anglo-Saxon
countries.
Another source of evidence is provided by tax amnesties, and
Alvaredo discusses the results for Argentina (Chapter 6). Information from
the 1962 tax amnesty (which attempted to uncover all income that had
been evaded by taxpayers between 1956 and 1961) suggested under-
reporting of between 27 and 40 per cent. However, it varied with income.
22
Evasion shows a lower impact at the bottom (where income from wage
source dominates) and at the top of the tax scale (where inspections from
the tax administration agency might be more frequent and enforcement
through other taxes higher). The evidence may be indirect. In the case of
India, Banerjee and Piketty (Chapter 1) note the innovations in tax
collection that may have affected the prevalence of filing. They investigate
the impact by considering the evolution of wage income, where taxes are
typically deducted at source, so that no change would be observed if all
that was happening was improved collection. They conclude that there was
a “real” increase in top incomes. As in other studies (such as that for
Australia in Volume 1, Chapter 7), this is corroborated by independent
evidence about what happened to top salaries.
It is important to remember that, while taxpayers may have a strong
incentive to evade, the taxing authorities have a strong incentive to enforce
collection. This takes the form of both sticks and carrots. For example, the
Inland Revenue Authority of Singapore devotes considerable resources to
enforcing tax collection, but also provides positive encouragement to tax
compliance through emphasising the role of taxes in financing key
government services such as schools. The resources allocated to tax
administration have been substantial: for example, in Spain in the pre-1960
period the administration was able to audit a very significant fraction (10-20
per cent) of individual tax returns. The tax authorities may also be
expected to target their enforcement activities on those with higher
potential liabilities. The scope for evasion may therefore be less for the
23
very top incomes than for those close to the tax threshold, as Leigh and van
Eng note to be the case in Indonesia (Chapter 4).
One important route to avoiding personal income tax is for income to
be sheltered in companies. The extent to which this is possible depends on
the personal tax law and on the taxation of corporations. One key feature is
the extent to which there is an imputation system, under which part of any
corporation tax paid is treated as a pre-payment of personal income tax.
Payment of dividends can be made more attractive by the introduction of an
imputation system, as in the UK in 1973, Australia in 1987 and New Zealand
in 1989, in place of a “classical” system where dividends are subject to both
corporation and personal income tax. Insofar as capital gains are missing
from the estimates (as discussed above) but dividends are covered, a switch
towards (away from) dividend payment will increase (reduce) the apparent
top income shares. This needs to be taken into account when interpreting
the results. That is why estimating series including realized capital gains is
valuable in order to assess the contribution of retained profits of
corporations on top individual incomes. When realized capital gains are
untaxed and hence not observed, it is important to assess the effects of
attributing retained profits to top income. For example, in the UK, Atkinson
(Chapter 4 in Volume 1) examined the consequences of the large increase
after the Second World War in the proportion of profits retained by
companies. The attribution of the retained profits to top income groups
would have reduced the magnitude of the fall in the share of the top 1 per
cent between 1937 and 1957 but still left a very considerable reduction.
24
The reported shares of top incomes can also be affected by the move
between incorporated and non-incorporated activities. This has been
modelled by Gordon and Slemrod (2000) and others. The potential impact is
particularly marked in the case of the dual income tax introduced in Nordic
countries. The tax reform in Finland in 1993 combined progressive taxation
of earned income with a flat rate of tax on capital income and corporate
profits, with a full imputation system applied to the taxation of distributed
profits. Under the dual income tax, capital income is taxed at a lower rate
than the top marginal tax rate on labour income. As discussed in Chapter 8,
the 1993 tax reform led to an increasing trend of the share of capital
income (dividends) and declining share of entrepreneurial income. This can
be interpreted as an indication of a tax induced shift in organisational form
and the choice of tax regime. In Chapter 10, Alvaredo and Saez provide a
model of the incentive to adopt a (wealth tax) exempt organisational form
and examine the effect of the wealth tax reform undertaken in Spain. Their
empirical estimates suggest that there is a very large shifting effect: the
fraction of businesses benefiting from the exemption jumps from 1/3 to
about 2/3 for the top 1%.
An extreme form of adjustment to income taxation is to leave the
country. In their study of Switzerland (Chapter 11 in Volume 1), Dell,
Piketty and Saez investigated the issue of tax evasion by foreigners
relocating to that country or through Swiss bank accounts. They found that
the fraction of taxpayers in Switzerland with income abroad or non-resident
taxpayers had increased in recent years but remains below 20 per cent even
at the very top of the distribution, suggesting that the migration to
25
Switzerland of the very wealthy is a limited phenomenon. They similarly
conclude that the amount of capital income earned through Swiss accounts
and not reported is small in relation to the total incomes of top income
recipients in other countries. In the case of Sweden, Roine and
Waldenström (Chapter 7) make estimates of “capital flight” since the early
1980s using unexplained residual capital flows (“net errors and omissions”)
published in official balance of payments statistics. They estimate that
somewhere between 250 and 500 billion SEK has left the country without
being accounted for. To get a sense of the order of magnitude by which this
“missing wealth” would change top income shares in Sweden, they add all
of the returns from this capital first to the incomes of the top decile and
then to the top percentile. For the years before 1990, there is no effect on
top income shares by adding income from offshore capital holdings since
they are simply too small. However, after 1990, and especially after 1995,
when adding all of them to the top decile, income shares increase
moderately (by approximately 3 percent). When instead adding everything
to the incomes of the top percentile, the income shares increase by about
25 percent which is equivalent to an increased share from about 5.7 to 7.0
percent. While this is a notable change, it does not raise Swedish top
income shares above those in France (about 7.7 percent in 1998), the U.K.
(12.5 percent in 1998) or the U.S. (15.3 percent in 1998).
To sum up, the different pieces of evidence indicate that tax evasion
and tax avoidance need to be taken seriously and can quantitatively affect
the conclusions drawn. They need to be borne in mind when considering the
results, but they are not so large as to mean that the tax data should be
26
rejected out of hand. Our view is that legally tax-exempt capital income
poses more serious problems than tax evasion and tax avoidance per se.
Summary
The data are rich but need to be used with due circumspection, particularly
with respect to incomes from capital. In drawing conclusions, users need to
ask themselves whether their findings could be reversed by taking into
account the inherent limitations of fiscal data, of the breaks in continuity
over time, and of the differences in methods that remain. Put differently,
there is a wide confidence interval surrounding the estimates, reflecting not
sampling error (since in many cases the statistics cover the universe) but
non-sampling error. In concrete terms, the different considerations
described above suggest that an error margin of ± 20 per cent is not
unreasonable, although it could well be exceeded if the different errors
cumulate in the same direction. In what follows, we take ± 20 per cent as a
yardstick.
13.3 A SUMMARY OF THE MAIN FINDINGS
Our summary of the evidence begins in the middle of the twentieth century.
The first columns in Table 13.1 show the position in 1949 (1950).5 We take
this year as one for which we have estimates for all except 4 of the 22
countries (for Indonesia we have taken the 1939 estimate and for Ireland
that for 1943), and as one when most countries had begun to return to 5 In the case of New Zealand, we have used the estimates of Atkinson and Leigh
(2008, Table 1) that adjust for the change in the tax unit in 1953.
27
normality after the Second World War (for Germany and the Netherlands we
take 1950). Moreover, it was before the 1950-1 commodity price boom that
affected top shares in Australia, New Zealand and Singapore.
If we start with the top 1 per cent – the group on which attention is
commonly focused – then we can see from Table 13.1 that the shares of
total gross income are strikingly similar when we take account of the
possible margins of error. There are 18 countries for which we have
estimates. If we take 10 per cent as the central value (the median is in fact
around 10.8), then 12 of the 18 lie within the range 8 to 12 per cent (i.e.
with an error margin of ± 20 per cent). In countries as diverse as India,
Norway, France, New Zealand, and the US, the top 1 per cent had on
average between 8 to 12 times average income. Three countries were only
just below 8 per cent: Japan, Finland and Sweden. The countries above the
range were Ireland, Argentina and (colonial) Indonesia. The top 1 per cent
is of course just one point on the distribution. If we look at the top 0.1 per
cent, shown in Table 13.1 for 18 countries (Portugal replacing Finland), then
we find that again 12 lie within a (± 20 per cent) range around 3.25 per cent
from 2.6 to 3.9 per cent. Leaving out the 3 outliers at each end, the top 0.1
per cent had between 26 and 39 times the average income.
We also report in Table 13.1 the inverse Pareto-Lorenz coefficients
associated to the upper tail of the observed distribution in the various
countries in 1949 and 2005. In this table, and throughout this chapter, we
choose to focus the attention upon the inverted-Pareto-Lorenz “β”
coefficient rather than the standard Pareto-Lorenz “α” coefficient. Note
that there exists a one-to-one, monotonically decreasing relationship
28
between the α and β coefficients, i.e. β=α/(α-1) and α=β/(β-1) (see the
notes to Table 13.1B). The reasons for using the β coefficient are twofold.
First, as was noted by Piketty (Chapter 1 in Volume 1), the β coefficient has
arguably greater economic appeal, in that it measures the average income
of people above y, relative to y. It provides a direct intuitive measure of the
fatness of the upper tail of the distribution. Next, a higher β coefficient
means larger top income shares and higher top income inequality (while the
reverse is true with the more commonly used α coefficient), which
facilitates the presentation and discussion of the results. In practice, we
shall see that the β coefficient typically varies between 1.5 and 2.5: values
around 1.5 or below indicate low top income inequality, while values around
2.5 or below indicate high top income inequality. A value of 1.5 means
that people above a specified level have on average 50 per cent more
income; a value of 2.5 means that they have 150 per cent more income.
Coming back to 1949, we find that 10 of the 20 countries for which β
coefficient values are shown in Table 13.1 lie between 1.88 and 2.00 in
1949. Countries as different as Spain, Norway, the US and (colonial)
Singapore had Pareto coefficients that differed only in the second decimal
place. As of 1949, the only countries with β coefficients above 2.5 were
Argentina and India.
1949 is of interest not just for being mid-century, but also because
later years did not exhibit the degree of similarity described above. The
right hand part of Table 13.1 assembles estimates for 2005 (or a close year).
The central value for the share of the top 1 per cent is not too different to
that in 1949: 9 per cent. But we now find more dispersion. For the top 1
29
per cent, 9 out of 21 countries lie outside the range of ± 20 per cent.
Leaving out the 2 outliers at each end, the top 0.1 per cent had between 13
and 56 times the average income (in 1949 these figures had been 20 and
52). In terms of the β coefficients only 4 of the 22 countries had values
between 1.88 and 2.00. Of the countries present in 1949, five now have
values of β in excess of 2.5.
The post-war picture
There was in face considerable diversity of experience over the period from
1949 to the beginning of the 21st century. If we ask in how many cases the
share of the top 1 per cent rose or fell by more than 2 percentage points
between 1949 and 2005 (bearing in mind that two-thirds were in the range 8
to 12 per cent in 1949), then we find the 17 countries more or less evenly
divided: 6 had a fall of 2 points or more, 5 had a rise of 2 points or more,
and 6 had a smaller or no change. If we ask in how many cases the inverted-
Pareto-Lorenz β coefficient changed by more than 0.1, then this was true of
15 out of 20 countries in Table 13.1, with 12 showing a rise (a move to
greater concentration). Examination of the annual data for individual
countries in Tables 13A.1 to 13A.22 confirms that during the 50+ years since
1949 individual countries followed different time paths.
Can we nonetheless draw any common conclusions? Is it for example
the case that all were following a U-shape, and that the differences when
comparing 2005 and 1949 arise simply because some countries are further
advanced? Is the US leading the way, with other countries lagging? In Table
13.2, we summarise the time paths from 1949 to 2005 for the 16 countries
30
for which we have fairly complete data over this period for the share of the
top 1 per cent and top 0.1 per cent. In focusing on change, we are not
interested in small differences after the decimal points. The criterion
applied in the case of the share of the top 1 per cent is that used above: a
change of 2 percentage points or more. For the share of the top 0.1 per
cent, we apply a criterion of 0.65 percentage points (i.e. scaled by
3.25/10). In applying this, we consider only sustained changes. This means
that we do not recognise changes due to tax reforms that distort the
figures, as in the case of Norway (Chapter 9) or New Zealand (see volume 1,
chapter 8), those due to the commodity price boom of the early 1950s, as
for Australia, New Zealand and Singapore, or other changes that are not
maintained for several years.
Applying this criterion, there is just one case – Finland – where there
is a pattern of rise/fall/rise. The share of the top 1 per cent in Finland rose
from below 8 per cent in 1949 (it has been lower before then) to around 10
per cent in the early 1960s. Of the remaining 15 countries, one can
distinguish a group of 6 “flat” countries (France, Germany, Switzerland,
Netherlands, Japan, Singapore), and a group of 9 “U-shaped” countries (UK,
US, Canada, Australia, New Zealand, India, Argentina, Sweden, Norway).
Broadly the same story is revealed by the β coefficients plotted in Figures
13.3 and 13.4 for both groups of countries.
The 10 countries belonging to the second group appear to fit, to
varying degrees, the U-shape hypothesis that top shares have first fallen and
then risen over the post-war period. In most countries, the initial fall was
of limited size, with β coefficients declining from about 1.7 in 1949 to about
31
1.5 during the 1970s, before climbing towards 2-2.5-3 during the 1990s-
2000s. In Argentina and India, there was higher concentration to start with
(β coefficients above 2.5 in 1949), and the declining part of the U-curve was
more marked (see Figure 13.4). The individual country patterns differ in
other respects as well. As may be seen from Table 13.2, the initial falls in
top shares were less marked in the US, Canada and New Zealand than in the
UK, Australia and India. The share of the top 1 per cent was much the same
in the US and UK in 1949 but in the UK the share then halved over the next
quarter century, whereas in the US it fell by only a little over a quarter.
From Figure 13.4 we can see that the decline in the β coefficient reached
1.7 in the US in 1969, the same value as in the UK, but the latter went on to
decline to about 1.5 by the late 1970s. Norway and Sweden reached values
as low as 1.3-1.4.
The frontier between the U-shaped countries and the flat countries is
somewhat arbitrary and should not be overstressed. In France, after an
initial reduction in concentration, the coefficient hovered around 1.7 from
1960 to the late 1990s, but has begun to rise since the late 1990s. In Japan
and Singapore, the rebound in recent years is even more pronounced (see
Figure 13.3 and the top income shares series in the Appendix tables). The
only three countries with no sign of a rise in income concentration during
the most recent period, namely Switzerland, Germany and the Netherlands,
are countries where our series stop in the late 1990s. There exists some
reasonable presumption that when data becomes available for the 2000s,
these countries might also display an upward trend. Finally, note that
Switzerland and especially Germany have always been characterized by
32
significantly larger concentration at the top than other Continental
European countries.
What about countries for which we have only a shorter time series?
The time series for China is indeed short, but there too the top of the
distribution is heading for greater concentration. For instance, β
coefficients have gradually risen from about 1.2 in 1986 to about 1.5 in 2003
(see Tables 13A.23 and 13A.24). These are still very small β coefficients by
international and historical standards, but the trend is strong (and the levels
might be under-estimated due to the nature of the available Chinese data,
see Chapter 2). China has a way to go, but the degree of concentration is
heading in the direction of the values in OECD countries. Regarding the
other countries with limited time coverage (Spain, Portugal, and Italy), one
also observes a significant rise in income concentration during the most
recent period.
Before 1949
What happened in our 22 countries before 1949 may appear like pre-history
to some readers, but the experience may be relevant for several reasons.
The behaviour of the income distribution in today’s rich countries may
provide a guide as to what can be expected in today’s fast-growing
economies. We can learn from nineteenth century data, such as those for
Norway, that cover the period of industrialization. Events in today’s world
economy may resemble those in the past. If we are concerned as to the
distributional impact of recession, then there may be lessons to be learned
from the 1930s.
33
The data assembled here provide evidence about the interwar period
for 19 of the 22 countries; and for 5 of the countries we have more than one
observation before the First World War. In Table 13.3 we have assembled
the changes in the shares of the top 1 per cent and top 0.1 per cent for
certain key periods, such as the world wars, and the crash of 1929-32, as
well as for the whole period up to 1949.
The first striking conclusion is that the top shares in 1949 were much
lower than thirty years earlier (1919) in the great majority of countries. Of
the 18 countries for which we can make the comparison with 1919 (or in
some cases with the early 1920s), no fewer than 13 showed a strong decline
in top income shares. In only 1 case (Indonesia) was there an increase in the
top shares. In half of the countries, the fall caused the shares to be at least
halved between 1919 and 1949. For countries where one can compare 1949
with 1913-1914, the fall generally seems at least as large.
What happened before 1914? In 5 cases, shown in italics, we have data
for a number of years before the First World War. Naturally the evidence has
to be treated with caution and has evident limitations: for example, the
German figures relate only to Prussia (see Dell, 2008, for estimates for Baden,
Hesse, Sachsen and Württemberg). But it is interesting that in the two Nordic
countries (Sweden and Norway) the top shares seems to have fallen somewhat
at the very beginning of the 20th century, a period when they might have been
in the upward part of the Kuznets inverted-U. As is noted in Chapters 7 and 9,
at that time Norway and Sweden were largely agrarian economies. In
neither Japan nor the UK is there evidence of a trend in top shares. (For the
German states the picture is less clear and varies across states – see Dell,
34
2008.) Given the scarcity of reliable income data for the pre-1914 period,
using wealth data is probably the most promising way to go in order to put
the First World War shocks into a long run historical perspective. Using large
samples of Parisian and national estate tax returns over the 1807-1994
period, Piketty, Postel-Vinay and Rosenthal (2006) have found that wealth
concentration rose continuously during the 1807-1914 period (with an
acceleration of the trend in the last three to four decades prior to 1914),
and that the downturn did not start until the First World War. Due to the
lack of similar wealth series for other countries, it is difficult to know
whether this is a general pattern. But for all countries where some pre-1914
evidence does exists, available information suggests that the sharp decline
in wealth concentration did not start before 1914 – or at least that the trend
was much more moderate prior to the First World War.
Are top incomes different?
In Volume 1, we emphasised the differences between the very top of the
distribution, the top 1 per cent, and the adjacent income recipients. In
Table 13.4 we assemble the findings for the “next 4 per cent” (those in the
second to fifth percentile groups) and the “second vingtile group” (those in
the sixth to tenth percentile groups). The values are shown for three of the
dates we have highlighted: around 1919 (or at the eve of the First World
War, when available), 1949 and 2005. We have added, in the final column,
text comments about these groups. In three cases, the data do not allow us
to estimate shares below that of the top 1 per cent, so that there are 19
countries shown.
35
In many cases – 15 out of 19 - the top 1 per cent are different, in the
sense that the changes in income concentration have been particularly
affected this group. For some countries, the “next 4 per cent” exhibit some
of the same features as the top 1 per cent (as in the UK in recent decades),
so that it would be fairer to talk of concentration among the top 5 per cent,
but typically the second vingtile group does not share the same experience.
In other cases, like China, it is a matter of degree. But this is not
universal, and in Table 13.4 we have shown in italics the 4 cases (Germany,
Japan, Singapore and Portugal) where there have been changes in the next
4 per cent and below.
Being in the top 1 per cent does not necessarily imply being rich, and
there are also marked differences within this group. The very rich are
different from the rich. We have earlier considered the top 0.1 per cent (in
Table 13.1), and a number of the chapters examine the top 0.01 per cent.
In Chapter 1, Banerjee and Piketty show that in India in the 1990s it was
only the top 0.1 per cent who enjoyed a growth rate of income faster than
that of GDP per capita, in contrast to the situation in the 1980s when there
was faster growth for the whole top percentile.
Composition of top incomes
In their study of the United States, Piketty and Saez found that the “rise in top
incomes is due not to the revival of top capital incomes, but rather to the very
large increases in top wages (especially top executive compensation). As a
consequence, top executives (the ‘working rich’) replaced top capital owners
(the “rentiers”) at the top of the income hierarchy during the twentieth
36
century” (2006, page 204). In France (Piketty, 2003), the top capital incomes
had not been able to recover from a succession of adverse shocks over the
period 1914 to 1945; progressive income and inheritance taxation had
prevented the reestablishment of large fortunes.
Data on the composition of top incomes are only available for around
half of the countries studied here, but a number record the decline of capital
incomes and the rise of top earnings. The Japanese data show that “the
dramatic fall in income concentration at the top was primarily due to the
collapse of capital income during the Second World War” (Moriguchi and Saez,
Chapter 3, page xx). In the Netherlands, “capital and wage incomes have
traded places within the top shares [although] the increased role of the latter
has not been able to prevent the decline or the stability of the top shares”
(Salverda and Atkinson, Chapter 10 in Volume 1, page 452). In Canada, “the
income composition pattern has changed significantly from 1946 to 2000. …
the share of wage income has increased for all groups, and this increase is
larger at the very top. ... The share of capital income [excluding capital gains]
has fallen very significantly for the very top groups” (Saez and Veall, Chapter
6 in Volume 1, page 239). The Italian data (Chapter 12) only start in 1974 and
the rise in top shares is modest: the share of the top 1 per cent rose from
around 7 per cent in the mid-1970s to around 9 per cent in 2004. But the
Italian data show a rise in the role of wage income in the very top groups. In
1976, earnings accounted for less than 10 per cent of the income of the top
0.01 per cent, but by 2004 this had increased to over 20 per cent. Over the
same period, the share of capital income more or less halved (Table 12A.4). In
Spain, a similar calculation (from figures that omit capital gains) shows that in
37
1981, earnings accounted for less than 20 per cent of the income of the top
0.01 per cent, but by 2004 this had increased to 40 per cent.
Further evidence can be obtained from other sources for some of the
countries without evidence on the composition of incomes in the income tax
data. In the case of Portugal, for example, the administrative records on
earnings show that the share of the very top earners increased: between 1991
and 2004 the share of the top 0.1 per cent doubled (Table 11D.6).
At the same time, the picture is not totally uniform. A major difference
between the Nordic countries and the US is the continuing importance in the
former of capital income. In Sweden, Roine and Waldenström find that
“between 1945 and 1978 the wage share at all levels of top incomes became
more important … But in 2004 the pattern is back to that of 1945 in terms of
the importance of capital, in particular when we include realized capital
gains” (Chapter 7, page xxx). The conclusions reached regarding Finland
stress that “the main factor that has driven up the top one per cent income
share in Finland after the mid 1990s is in an unprecedented increase in the
fraction of capital income” (Chapter 8, page xxx). This may reflect
differences in reporting behaviour following tax reforms, but it is not totally a
difference between Nordic countries and the Anglo-Saxons. In Australia,
Atkinson and Leigh found that “the proportion of salary and wage income for
top income groups in 2000 was quite similar to the proportion in 1980”
(Chapter 7 in Volume 1, page 322). In the UK, it is true that the major themes
have been the fall in capital incomes over the first three-quarters of the 20th
century and the subsequent rise in top earnings, but minor themes have been
38
an earlier fall on the share of top earners and a partial restoration of capital
incomes since 1979.
Summary
It is not easy to summarise a summary, not least because to almost every
statement there is a counterexample among the 22 countries we have been
studying.
• At the middle of the 20th century, the top of the income distribution
looked similar in many of the different countries for which we have
data: for two thirds the top 1 per cent had on average between 8 and
12 times average income. Countries as different as Spain, Norway,
the US and (colonial) Singapore had inverse Pareto coefficients that
differed only in the second decimal place.
• This was to change: from 1950 to the present, countries followed
different paths, and there is now greater diversity. Out of 17 countries
we can track from 1949 to 2005, 6 had over the period as a whole a
fall of 2 points or more in the share of the top 1 per cent, 5 had a
rise of 2 points or more, and 6 had a smaller or no change.
• Within the period, the majority of countries appear to fit, to varying
degrees, the U-shape hypothesis that top shares have first fallen and
then risen over the post-war period. This was not universal: a
number exhibited either no change or a limited recent rise in top
shares.
39
• The post-war fall in top shares (where it happened) may be seen as
continuing the pattern of the first half of the century, but the 1900-
1945 period was particularly affected by events, including, for both
combatants and non-combatants, the two world wars, and the Great
Crash of 1929.
• In most countries, the changes, and particularly the recent increases,
have been concentrated at the very top.
• The decline in top income shares over the first three-quarters of the
20th century was largely associated with a decline in top capital
incomes; the recent rise in top shares in a number of countries has
been particularly associated with increased top earnings, but this is
not universal and in the Nordic countries the rise was associated
largely with capital income.
One way to summarize our findings over the entire 1900-2005 period is again
to plot separately top income shares and β coefficients for the two groups of
countries defined above, which now become the “L-shape” group and the “U-
shape” group – keeping in mind that the frontier between both groups is fuzzy,
and that L-shape countries seem to be gradually shifting towards the U-shape
pattern (see Figures 13.5, 13.6, 13.7 and 13.8).
13.4 SEEKING POSSIBLE EXPLANATIONS: EMPIRICAL AND THEORETICAL
MODELS
40
From the data on the changes in the upper part of the income distribution
assembled in these two volumes certain possible explanations stand out. We
have drawn attention to the falls in top income shares in countries fighting
in the First and Second World Wars (and that some, but not all, non-
combatant countries, were less strongly hit, or even saw an increase in top
shares). According to Moriguchi and Saez, “the defining event for the
evolution of income concentration in Japan was a historical accident,
namely the Second World War” (Chapter 3, page xx). Much less momentous,
but still distinctive as an event, was the commodity price boom of 1950,
which saw a rise in top shares in Australia, New Zealand, and Singapore. In
these cases, a single event is sufficiently large for us to be content with a
single variable analysis. Moreover, there is unlikely to be reverse causality,
with the fall or rise in shares causing the wars or the commodity boom. The
slump in commodity prices in the Great Depression may also be such an
event – see the discussion of Indonesia by Leigh and van der Eng in Chapter 4
– but the wider economic circumstances were also highly relevant.
Indeed, in general, explanations are likely to be multivariate, and we
are confronted with the task of seeking to separate different influences. In
the introductory chapter to Volume 1, Piketty suggested that the database
constructed here could be exploited as a cross-country panel, and this
approach has already been adopted by Roine, Vlachos and Waldenström
(2008) and Atkinson and Leigh (2007). The former authors find, for example,
that growth in GDP per head is associated with increases in top income
shares and that financial development is pro-rich in the early stages of a
country’s development.
41
Multivariate statistical analysis may help us disentangle some of the
factors at work. For example, a number of the chapters, following Piketty
(2001 and 2003) highlight the role of progressive income taxation. In the UK,
for example, the period of falling top income shares was one of high
marginal rates of tax, and the shares began to rise again when the tax rates
were sharply reduced in the 1980s. But how can we be sure that there is a
causal path from progressive taxation to reduced top income shares? In the
UK, high top rates of income tax were first introduced during the First World
War. Could these tax rates, and the reduction in top shares, not be seen as
both resulting from third factors associated with the War and its aftermath,
such as the loss of overseas income? Statistical analysis seeks to separate
out the independent variation in different variables. For example, the UK
was a combatant in the First World War but the Netherlands (also a colonial
power) was not. It may therefore be informative to compare the 2
countries, both of which had progressive income taxes. At the same time,
there are possible third factors. Both the UK and the Netherlands faced
similar global economic conditions that may have independently affected
top shares. In the same way, the tax cuts of the 1980s took place under
Reagan and Thatcher, just as the First World War increases in the UK had
been initiated by Liberal Governments. These governments pursued other
policies apart from income taxation, such as the measures to prevent
profiteering in the First World War, or the liberalisation of the capital
markets and privatisation in the 1980s, which may have affected the top
income shares. There is also the possibility of reverse causality. The
increases in top incomes as a result of changed executive remuneration
42
policies may have increased political pressure for cutting top taxes. We
need therefore a simultaneous, as well as multivariate, model.
Statistical analysis can help us identify independent variation, but it
rarely proves fully conclusive. The conclusions that we draw inevitably
involve elements of judgment. Judgment in turn is likely to be influenced by
a range of considerations. Here we consider two: historical narrative and
economic theory (on which we particularly focus).
The conclusion regarding the role of progressive income taxation in
France was reached by Piketty after an extensive discussion of the economic
history of France over the twentieth century. While it would be re-inforced
by regression analysis in which the relevant tax rate variable had a highly
(statistically) significant coefficient of a plausible magnitude, the conclusion
was based on a reading of the events of the period. In the same way, the
individual chapters in these two volumes provide each a historical narrative
that in itself is part of the evidence. The narrative typically draws on a
variety of evidence. A number of chapters, such as that on Japan, contain
evidence from a range of sources: income tax data, wealth data, estate
data, and wage data. In combining these disparate sets of information, the
authors are not carrying out a mechanical operation, but exercising
judgment about the strengths and weaknesses of different sources. These
narratives are of course subjective, reflecting the standpoints of the
authors, and there will no doubt be disagreement about the interpretation
of history. Again they cannot be definitive. But equally they cannot be
dismissed out of hand, and they play a significant role in our summary of
major mechanisms in the next section.
43
Theoretical models
The judgment concerning the importance of progressive taxation in France
was also reached on the basis of theoretical considerations, notably
simulation models of capital accumulation. This brings us to the question of
the relation between theoretical models of income distribution and the
specification of statistical relationships. How closely are these to be linked?
In the theory of consumer demand, if a consumer maximises a logarithmic
utility function (of the consumption of each good plus a constant as in the
Stone-Geary form) subject to a linear budget constraint, we can derive the
predicted demands as linear functions of income. This straight line curve
can be fitted to expenditure data, and since the days of Engel this has
provided a valuable framework for understanding how consumer spending is
likely to change over time. The coefficient on income is interpreted as the
marginal propensity to consume, and the specified functional form allows
inferences can be drawn about the response to price changes. There is in
this case a tight link between the theoretical model and the empirical
implementation.
In contrast, the income inequality literature has typically a looser
connection (see Atkinson and Brandolini, 2006 for a survey). Theoretical
models are invoked, but to produce a list of explanatory variables rather
than to generate an estimating equation. The functional form is not
specified, so that it is not clear how the explanatory variables should enter
the estimating equation. Should the model be linear? Should the
explanatory variables interact? There is no guide to the form of the variable
44
to be explained. Should the left hand side be the top income share? Should
it be a transformation? Should it be the (inverse) Pareto coefficient?
Modelling sectoral shifts
Building a link between theory and empirical specification is not
straightforward, as may be illustrated by reference to the most popular
model in the income distribution literature: the Kuznets inverse-U curve.
This curve is based on the structural change that takes place in an economy
as it is transformed from largely agricultural (traditional) to industrial
(modern). We should note, before using this model, that its popularity
seems to far exceed its demonstrated empirical relevance. As stressed by
Piketty (2001 and 2003), in the first post-Kuznets study of top incomes, the
inverse-U has little purchase in explaining top income shares; indeed, he
argues that we should look instead at those sections of Kuznets (1955)
article where he emphasises the factors counteracting the concentration of
savings, notably the impact of progressive taxation – to which we return
below.
If we take the Kuznets model of structural change, what does it imply
for top income shares? With his numerical assumptions, the top income
group is the top decile of those in the higher-paid industrial sector. They
initially constitute, when the agricultural sector employs 80 per cent of the
population, 2 per cent of the population. Half their share (that of the top 1
per cent) in the Kuznets numerical example is either 2.4 per cent
(moderately unequal) or 3.2 per cent (more unequal) of total income. As the
industrial sector grows, this share falls: with the agricultural sector
45
employing 70 per cent of the population, the shares become 2.2 and 2.9 per
cent. With a linear top section of the Lorenz curve (as assumed in Kuznets’
example) , the share of the top 1 per cent, S1, is simply a constant divided
by mean income (and hence is strictly decreasing with mean income). (More
realistic would be an assumption that the upper tail of incomes in the
industrial sector follows some distribution such as the Pareto.)
The basic problem with the Kuznets model as far as top shares are
concerned is that it focuses essentially on labour income, whereas it is clear
that we need to consider both labour and capital income, and their changing
roles. Indeed it is with capital incomes that we start, since historically they
accounted for the bulk of top incomes.
Modelling capital incomes
In the first part of his Presidential Address, Kuznets (1955) evokes two
“groups of forces in the long-term operation of developed counties [that]
make for widening inequality in the distribution of income” (1955, page
xxx). The first of these is the concentration of savings in the upper income
brackets and the cumulative effect on asset holding. Subsequently, Meade
(1964) developed a theory of individual wealth holding, allowing for
accumulation and transmission of wealth via inheritance, and this model has
been analysed in a general equilibrium setting by Stiglitz (1969). With equal
division of estates at death, a linear savings process, and persistent
differences in earnings across generations, in the long-run the steady-state
distribution of wealth simply mirrors the distribution of earnings (Atkinson
and Harrison, 1978, page 211). If the society starts with a more unequal
46
distribution of capital than steady-state level, then the top shares will fall
in the approach to this equilibrium (and conversely if it starts below steady-
state levels, say if earnings inequality increased for some exogenous
reason). But to explain the extent of inequality we have to have appeal to
explanations of the distribution of earnings.
Alternative assumptions about bequests can however generate long-
run equilibria where there is inequality of wealth even where earnings are
equal. Stiglitz shows how the operation of primogeniture (leaving all wealth
to one child) can lead in equilibrium to a stable distribution with a Pareto
upper tail, with the Pareto coefficient
α = log[1+n] / log[1+sr(1-t)] (1)
where sr(1-t) is the rate of accumulation out of wealth, s being the savings
rate, r being the rate of return, t the tax rate, and n is the rate of
population growth (Atkinson and Harrison, 1978, page 213). For stability,
the population growth rate has to exceed the rate of accumulation by the
wealthy, so it follows that α is greater than 1. The faster the rate of
accumulation, the closer α is to 1. Equation (1) provides a – deceptively –
simple answer to the questions concerning specification. Approximating
log(1+x) by x, we should regress 1/α (or 1/β) on sr(1-t)/n. This provides a
natural way of testing the impact of progressive income taxation. However
this is indeed deceptive, since it assumes that the parameters are constant
over time, and that the primogeniture assumption is remotely plausible.
The first of these concerns might be met by using a moving average of past
tax rates. In countries such as the UK where the top tax rate was cut from
47
98 per cent to 40 per cent in the first half of the 1980s, there would then be
a continuing rise in top shares until the new equilibrium was approached.
Wealth is not typically concentrated in a single line of descent.
Primogeniture may have applied in aristocratic England, but it was not
legally permissible in most European countries (and, after 1947, Japan) and
it never became widely established in the United States. With equal division
of inheritance, which seems a more reasonable assumption, Vaughan (1993)
has shown that the equation for the equilibrium value of α is given by the
implicit form:
α[sr(1-t)-½ σ2r2(1-t)2] + ½σ2r2(1-t)2α2 = (δ+n)[1-(1+n/δ)-α] (2)
where δ is the rate of mortality. In this model, Vaughan has also introduced
a random element in the return to capital (and for the underlying portfolio
choice), where σ2 is the variance of the white noise stochastic process. If
there is no randomness, and n = 0, so that the population is not growing (so
primogeniture is the same as equal division), then we have (1/α) = sr(1-t)/δ,
similar to the earlier estimating equation. If n is small relative to δ, we can
approximate the power on the right hand side of (2) and solve for α as a
decreasing linear function of the net rate of return, r(1-t), and which falls
with the contribution of the stochastic term, σ2r2(1-t)2.
The models of top incomes described above relate to capital income;
we need now to consider possible explanations in terms of earned incomes.
Modelling top earnings
Referees of a number of the papers in these two volumes, when they were
submitted for journal publication in an earlier form, took the authors to task
48
for not paying sufficient attention to the dominant paradigm in labour
economics, which explains rising wage dispersion in terms of skill-biased
technical change. Or, as it had earlier been put by Tinbergen (1975), there
is a race between the expansion of education and the increased demand for
educated labour as a result of technological change. While we agree that
this literature offers important insights, we do not feel that it has a great
deal to say about what is happening at the very top of the earnings
distribution. Empirically, labour economists have discussed the top decile as
a proportion of the median, but we are interested in what happens to the
top percentile and within the top percentile group. The skill-bias
explanation has highlighted the premium to college education (see, for
example, Katz and Autor, 1999), but that has little to say directly about why
the top percentile has increased relative to the top decile. The recent
“polarisation” thesis of Autor, Katz and Kearney (2006) has specifically
focused on the impact of technical change in replacing routine manual jobs,
which only indirectly affects top earners.
There are in fact a number of earlier theories that are directly
relevant to top earnings. One such set of theories are those dealing with
executive remuneration in a hierarchical structure. The model advanced by
Simon (1957) and Lydall (1959) generates an approximately Pareto tail to
the earnings distribution, with a Pareto exponent given by
α = log[span of managerial control]/log[1+ increment with promotion] (3)
49
In this form, the model is purely mechanical, but it provides a vehicle by
which we may introduce a number of explanatory variables, including
technological change, taxation, and changes in the size distribution of firms
and other organisations. Tournament theory (Lazear and Rosen, 1981), for
example, has provided an explanation of the size of the necessary
increment. If one considers the position of people at particular level in an
organisation, deciding whether or not to be a candidate for promotion to
the next rank, then they are comparing the certainty of their present
position with the risk of taking a new position in which they may fail, and
lose their job. The higher rank job also involves greater effort. In the very
simplest case, the worker weighs the mean gain against the risk. There are
two competing effects. On the one hand, the tax reduces the financial gain
from promotion and more is needed to compensate for the increased effort.
On the other hand, the tax reduces the risk of the new job: the government
shares part of the risk.
A second explanation of the rise in top earnings shares in a number of
countries in the second half of the post-war period is provided by the
"superstar" theory of Rosen (1981). The expansion of scale associated with
globalisation and with increased communication opportunities has raised the
rents of those with the very highest abilities. Where the “reach” of the top
performer is extended by technical changes such as those in Information and
Communications Technologies (ICT), and by the removal of trade barriers,
then the earnings gradient becomes steeper. Moreover, Frank and Cook
(1995) argue that the winner-take-all payoff structure has spread beyond
fields like sport and entertainment: “it is fair to say that virtually all top-
50
decile earners in the United States are participants in labour markets in
which rewards depend heavily on relative performance” (Frank, 2000, page
497). This could explain the fall in the Pareto α coefficient (and the rise in
the β coefficient) in the past quarter century. Indeed Rosen made precisely
this prediction in 1981, referring back to Marshall’s Principles, where
Marshall identifies “the development of new facilities for communication,
by which men, who have once attained a commanding position, are enabled
to apply their constructive or speculative genius to undertakings vaster, and
extending over a wider area, than ever before” (1920, page 685). As
captured in the title of the book by Frank and Cook (1995), it is a Winner-
Take-All Society, and this suggests that it can usefully be modelled as an
extreme value process. The distribution of earnings in this case is given by
the maximum values generated by the results of many separate
“competitions”. If we limit attention to those values exceeding some
specified threshold, then for a sufficiently high threshold the distribution
function takes on the generalised Pareto form (Embrechts, Klüppelberg and
Mikosch, 1997, page 164 or Coles, 2001, page 75), which has a Pareto upper
tail.
Finally, considerable attention has been devoted to the effects of
marginal tax rates--and especially top marginal tax rate--on the earnings
distribution. Higher top marginal tax rates can reduce top reported
earnings through three main channels. First, top earners may work less and
hence earn less—the classical supply side channel. Second, top earners may
substitute taxable cash compensation with other forms of compensation
such as non-taxable fringe benefits, deferred stock-option or pension
51
compensation—the tax shifting channel.6 Third, because the marginal
productivity of top earners, such as top executives, is not perfectly
observed, top earners might be able to increase their pay by exerting effort
to influence corporate boards. High top tax rates might discourage such
efforts aimed at extracting higher compensation.7
The central concept capturing all those behavioural responses to
taxation is the elasticity of reported earnings with respect to the net-of-tax
rate (defined as one minus the marginal tax rate). There is a large literature
(surveyed in Saez, Slemrod, and Giertz, 2009) which attempts to estimate
this elasticity. In general, the literature estimate this elasticity based on
the sum of labor and capital income although, as we discussed above, the
effects of tax rates on capital income might have a fairly long lag.
With a constant and uniform elasticity e, and a marginal tax rate t,
by definition, reported earnings will be: z=z0(1-t)e ,where z0 is reported
income when the marginal tax rate is zero. Therefore, the top income share
will be proportional to (1-tT)e/(1-tM)e
where tT is the top group marginal tax rate on earnings and tM is the average
marginal tax rate on earnings. Therefore, top income shares, combined with
information on marginal tax rates by income groups can be used to test this
6 The taxation of stock-options varies substantially across countries, In the United States, profits from stock-option exercises are included in wages and salaries for tax purposes and hence captured in the estimates. In other countries, such as France, profits from stock-options are taxed separately and hence are not included in the estimates. 7 The welfare consequences of taxation differ widely across the three channels. The first channel creates pure tax distortions. In the second channel, the tax distortion is reduced by “fiscal externalities” as tax shifting might generate deferred tax revenue as well. In the third channel, taxes can actually correct a negative externality if the contract between the executive and the board does not take into account the best interests of shareholders and other wage-earners.
52
theory and estimate the elasticity e with a log-form regression specification
of the form:
log(Top Income Share) = α + e log(1-tT) + ε,
As discussed below, Saez (2004) proposes such an exercise with US data
from 1960 to 2000. Atkinson and Leigh (2007) and Roine, Vlachos, and
Waldenström (2008) combine data from several countries (and include
several other variables) to test this relationship. In all those studies, top
marginal tax rates do seem to negatively affect top income shares, although
causality is difficult to establish. The main limiting factor to extend such an
analysis is the absence of systematic series on marginal tax rates by income
groups.8
Combining capital and earned income
In order to explain the shifting mix of capital and earned income, we need
to bring the two income sources together in a single model. This crucially
depends on their joint distribution. Are those with large capital incomes
also those with high salaries, accumulating assets over their careers? Or are
there, as assumed in classical distribution theories, separate classes of
“workers” and “capitalists”?
The latter case, with two distinct groups with high incomes, is the
easier to handle. We can consider the upper tail of the income distribution 8 Top marginal income tax rates may not approximate well effective marginal tax rates in upper income groups because of various exemptions, special provisions, the presence of other taxes such as social security contributions or local income taxes. When top tax rates were extremely high, the fraction of taxpayers in top bracket was often extremely small as well so that the marginal tax rate in the top 1% was substantially lower than the top marginal tax rate.
53
being formed as a mixture of the two upper tails. For example, if the upper
tail of earnings, for reasons we have just discussed, has a Pareto relative
cumulative distribution with exponent αl, and capital income is distributed
according to a Pareto distribution with exponent αk, then the overall
distribution could be seen as a combination of the two: a simple mixture.
The shape of the cumulative distribution depends on the relative weight on
the two distributions, and in this way we can introduce the overall shares of
wage and capital income (factor shares). If the exponent is assumed to be
less for capital income than for earnings (i.e. αk < αl), then those with
capital income become increasingly dominant as we move up the income
scale.
Where people receive both earned and capital income, we have to
make assumptions about their correlation. Where they are independent, we
have the convolution of the two distributions. This again introduces the
relative shares of earned and capital income in total income. However, this
approach does not offer any obvious simple functional forms (since we are
adding not multiplying the two components). Moreover, it seems more
realistic to assume some positive degree of correlation. In the extreme case
where people are ranked the same in the two distributions, we can form the
combined distribution by inverting the cumulative distribution. Expressing y
as a function of (1-F), we have in the case of the Pareto distribution, y =
[A/(1-F)]1/α. So that, if we add earned and capital income, we have total
income as
[A/(1-F)]1/αl + [B/(1-F)]1/αk (4)
54
Where αk < αl , the ratio of capital to earned income rises as we move up
the distribution. Again the relative weight of capital and labour income
enters via the constants A and B.
The different elements may be brought together in a simple
decomposition. Taking for illustration the share of the top 1%, this can be
broken down as follows:
Share of top 1% =
Proportion of earned income x Share of top 1% of earners
x Alignment coefficient for earnings
+
Proportion of capital income x Share of top 1% with capital income
x Alignment coefficient for capital income
(5)
The “alignment coefficient” for earnings (capital income) is the share in
earnings (capital income) of the top 1% of income recipients divided by the
share of top 1% of earners (capital income recipients). Since the top 1% of
earners (capital income recipients) are not necessarily in the top 1% of
income recipients, the alignment coefficient is by definition less than or
equal to 1. It is equal to 1 in the case discussed at the end of the previous
paragraph, but in a class model where no workers are in the top 1% the
coefficient is zero. Evidence about the degree of alignment in the case of
Sweden is provided in Chapter 7, which shows the distribution of wealth
both ranked by wealth and by total income. As may be seen from Figure 7.8,
the share in total wealth of the top 1 per cent is some 5 to 10 percentage
55
points lower when ranked by total income, but the two series move closely
together over time.
Summary
The above examples give some idea of the strength of assumptions that are
necessary to bridge the gap between theoretical models and empirical
specification. For some readers the assumptions required may indeed be a
bridge too far, and proof that we have simply to accept ad hoc
specifications. Other readers however may see the formulation as solid
ground in shifting sands, even if someway removed from where we would
like to be. Our view is that micro-based models, in particular micro-based
formulae for (inverse) Pareto coefficients, probably provide the most
promising strategy to develop convincing empirical tests of the determinants
and consequences of income and wealth concentration – probably more
promising than standard cross-country regressions. However our data set,
especially because of its lack of systematic decomposition between labour
income and capital income components, and of systematic series on labour
and capital tax rates, is unfortunately insufficient to do this in a fully
satisfactory manner at this stage.
13.5 SEEKING POSSIBLE EXPLANATIONS: MAJOR THEMES
In this section we consider some of the major explanatory factors suggested
by the theoretical models described in the previous section and by the
country accounts given in this volume and in volume 1.
56
Politics and political economy
The period covered by our top income data have seen great changes in the
political landscape. In 1900, all but 4 of the 22 countries studied here were
(or were ruled by) monarchies (the exceptions were Argentina, France,
Switzerland, and the US). Before the First World War, a quarter of the
world’s population lived as part of the British Empire. When the League of
Nations was founded in 1920, there were just 42 member countries. Of the
22 countries studied, six gained their independence since 1900. Many of the
countries saw significant changes in their boundaries, such as the partition
of India, and the division and re-unification of Germany. Most of the
countries were combatants in either the First or Second World Wars, and all
were affected by these wars. The countries studied here include 4 of the 6
that founded the European Union, and ten are current members of the EU.
In Table 13.5, we have summarised some of the main events that affected
the 22 countries during the period since 1900.
The most momentous events were the world wars, and for most
countries these were associated with falls in the top income shares. Starting
with the Second World War, for 14 countries we can observe the shares before
and after entry into the war. Of these, one showed an increase: Argentina,
where the top income shares were buoyed by expanded food exports to
combatant countries (see Chapter 6). The remaining 13 all saw the top shares
fall (for Germany no comparison is possible). The falls were again large: the
share of the top 0.1 per cent fell by a third or more in France, the US,
Canada, the Netherlands, Japan and Norway. For the First World War, we
57
have fewer observations. The top shares rose in the Netherlands, which was a
non-combatant, but they fell in all of the 3 combatants in Table 13.3 for
whom data exist: Japan, the UK and the US.
What caused the falls in top shares during world wars? Two forces
seem to have been in operation. The first, and probably much the most
important, was the loss of capital income. For France, Piketty stresses that
“the physical destructions induced by both World Wars were truly enormous
in France. … about one-third of the capital stock was destroyed during the
First World War, and about two-thirds during the Second World War”
(volume 1, page 56). This was followed in 1945 by nationalisation and a
capital levy. The UK lost during the wars much of its capital income from
abroad. In 1910 UK net property income from abroad represented 8 per cent
of GNP, by 1920 it had fallen to 4½ per cent; in 1938 it was close to 4 per
cent, but by 1948 it had fallen to under 2 per cent (Feinstein, 1972, Table
1). In the case of Japan, Moriguchi and Saez attribute the precipitous fall in
income concentration during the Second World War primarily to the collapse
of capital income due to wartime regulations, inflation and wartime
destruction. They go on to argue that the change in the institutional
structure under the Allied occupational reforms made the one-time income
de-concentration difficult to reverse. The reductions in capital incomes also
reflected the rise in corporate taxes during the wars and the restrictions on
the payment of dividends.
The second mechanism by which world wars led to falls in top shares
is via an equalisation of earned incomes. In the US, Goldin and Katz (1992)
have applied the term “the Great Compression” to the narrowing in the
58
United States wage structure in the 1940s: “when the United States
emerged from war and depression, it had not only a considerably lower rate
of unemployment, it also had a wage structure more egalitarian than at any
time since.” (1992, page 2). The war economy imposed wage controls,
under the National War Labor Board, as described by Piketty and Saez in
Chapter 5 of volume 1. Saez and Veall find that a compression also took
place in Canada during the war years (Chapter 6 in volume 1). In Japan, the
share in total wages of the top 5 per cent fell from 19 per cent in 1939 to 9
per cent in 1944 (Table 3C.2).
Along with wars went changes in political regimes, either as a
consequence or as a cause. The countries studied include five that were
governed by dictatorships/military rule during the period covered by our
data: Argentina, Germany, Indonesia, Portugal and Spain. It is not possible
in all cases to use the top income series to investigate their distributional
impact, since the dictatorship coincided with the virtual absence of data
(Argentina and Indonesia). But for some countries conclusions can be
drawn. Of Germany, Dell writes: “when the Nazis came to power in 1933,
the top decile had been thoroughly equalized … The effect of Nazi economic
administration changed radically this outcome … In a period of time of only
five years, the pre-First World War shares were nearly recovered” (Chapter
9 in volume 1, page 374). In contrast, in the case of Spain, Alvaredo and
Saez (Chapter 10) find that the top income shares fell during the first
decade of the Franco dictatorship. They also conclude that the transition
from dictatorship to democracy was not associated with a significant change
in top shares. This latter finding in turn may be contrasted with that for
59
Portugal, where Alvaredo finds a downward jump in top shares after 1970,
and particularly 1974. He notes that this “coincided with the final period of
the dictatorship and could be attributed to the loss of the African colonies
and to the leftward movement of the revolutionary government after 1974,
when a process of nationalizations broke up the concentration of economic
power in the hands of the financial-industrial groups.” (Chapter 11, page
XXX).
Within democracies, the top shares may be affected by changes over
time in political partisanship – whether Clinton or a Bush was in the White
House. It is naturally tempting to relate the observed changes over time to
political variables. Scheve and Stasavage (2009) use a panel of top income
data for 13 countries, but cannot find any strong effect of partisanship. This
will doubtless be further explored. Political variables may be more relevant to
explaining differences across countries, reflecting political climate and
traditions. As is noted by Roine and Waldenström in Chapter 7, a distinction
is often drawn between liberal (Anglo-Saxon) welfare states, corporatist-
conservative (Continental European) welfare states, and social democratic
(Scandinavian) welfare states. This makes it interesting to compare top
income shares in Sweden and Norway with those in the US/UK and in France
and Germany.
Finally, a major change in political regime is the end of colonial rule.
The 22 countries include three for which we have data before and after
independence. In the case of Indonesia, however, there is too large a gap in
time to draw conclusions. In India, as with Indonesia, independence
coincided with the end of the Second World War, so that it is hard to
60
distinguish the effect of independence per se. Only for Singapore do we
have observations for a post-war colonial period. Here, as shown in Chapter
5, there is little evidence of a decisive break in the top income series with
self-government.
61
Table 13.5 Summary of major political changes over period since 1900 for countries in Volumes 1 and 2. Country Main events (first observation)
Volume 1
France 1905
Combatant in First World War 1914-1918 Occupied during Second World War
UK 1908
Combatant in First World War 1914-1918 Combatant in Second World War 1939-1945
US 1913
Combatant in First World War 1917-1918 Combatant in Second World War 1941-1945
Canada 1920
Combatant in First World War 1914-1918 Combatant in Second World War 1939-1945
Australia 1921
Combatant in First World War 1914-1918 Combatant in Second World War 1939-1945
New Zealand 1921
Combatant in First World War 1914-1918 Combatant in Second World War 1939-1945
Germany 1896 (Prussia)
Combatant in First World War 1914-1918. Republic 1918 with reduced territory. Hitler Chancellor 1933. Combatant in Second World War 1939-1945. Occupied and Federal Republic 1949. Re-unified 1990.
Netherlands 1914
Occupied in Second World War
Switzerland 1933
Ireland 1922
Irish Free State 1922. Neutral in Second World War.
Volume 2 India 1922
Combatant in First World War 1914-1918 Combatant in Second World War 1939-1945 Partition and independence in 1947
China 1986
Japan 1886
Combatant in First World War 1914-1918 Combatant in Second World War 1941-1945 Occupied until 1952.
Indonesia 1920
Dutch colony. Occupied during Second World War. Independence in 1945. Military rule (Suharto) 1966-1998.
Singapore 1947
British colony. Internal self-government 1959. Joined Malaysia 1963. Expelled from Malaysia and fully independent from 1965.
Argentina 1932
Neutral in Second World War. Peron Presidency 1946, deposed in 1955 (brief return in
62
1974). Military coups d'état in 1930, 1943, 1955, 1962, 1966 and 1976.
Sweden 1903
Neutral in both world wars.
Finland 1920
After declaration of independence from Russia and civil war, Finland became a republic in 1919. Engaged in Winter War 1939-40, Continuation War 1941-44 and Lapland War 1944-45. Ceded around 10% of territory to Russia in treaty of 1947.
Norway 1875
Separated from Sweden in 1905. Neutral in First World War. Occupied in Second World War.
Spain 1933
Spanish Civil War 1936-9. Franco dictatorship. Neutral in Second World War. Democracy restored in 1976.
Portugal 1936
Salazar dictatorship. Neutral in Second World War. Democracy restored in 1974 following the peaceful “Carnation” revolution.
Italy 1974
63
Macro-economics and financial crises
Today there is much interest in looking back to the Great Depression. What
were the distributional consequences of major recession? Was it bad for top
income shares? Among the 13 countries for which we have data, the period
1928-31(2) saw a rise in top shares in Canada (top 1 per cent), India, Indonesia
and Ireland, and no change in Finland and Germany. The remaining 7 all saw
top shares reduced. The top 0.1 per cent lost a fifth or more of their income
share in Australia, France, the Netherlands, New Zealand, the UK, and the US.
In many countries, therefore, the depression reduced inequality at the top.
How far is this borne out by the historical accounts for individual
countries? For the US, Piketty and Saez (Chapter 5 in volume 1) find that the
share of the top 0.01 per cent fell sharply from 1929 to 1932, in the sense that
their average income went from 300 times the mean to 200 times. In the UK
the same group saw their average income fall from 300 to 230 times. In the
Netherlands, the top 0.05 per cent saw their share fall from 5.6 to 3.4 per
cent. In contrast, the fall in Japan in top shares was much smaller. In the
case of Sweden, Roine and Waldenström draw attention to the depression
hitting Sweden later in 1931 (although they note that the depression of the
1920s was more severe), and in particular the dramatic collapse of the
industrial empire controlled by the Swedish industrialist Ivar Kreuger in
1932. They show that between 1930 and 1935 there was a drop from 50
percent to 43 percent in the top percentile wealth share but an even larger
drop in the wealth of the top one percent of income earners, from 38
percent in 1930 to 26 percent in 1934.
64
1929, like 2008, combined the onset of a wide recession with a
financial crisis. What can we say about the latter from other episodes of
financial crisis? In the case of Norway, there are grounds for believing that
the Kristiana crash in 1899 led to a fall on top income shares (Chapter 9).
Much more recently, however, the Norwegian banking crisis of 1988-1992
does not appear to have led to a fall in top shares, although it may have
postponed the increases associated with financial market liberalisation. It is
possible that today’s financial crises are different from those in the past in
their distributional consequences. In the case of Singapore, top income
shares rose following the financial crisis of 1996-7, even if they have fallen
back to some extent subsequently. In Indonesia (Chapter 4), there are some
similarities.
Turning to the wider macro-economic determinants of top shares, we
saw in our discussion of the theoretical models, that an important role is
potentially played by the relative shares of earned and capital income.
These are related to, but not identical to, factor shares in GNP. As is shown
by Piketty for France (Figure 3.4 in volume 1), the capital share in
household income follows a different path from the corporate share in value
added. The same is demonstrated for the US by Piketty and Saez (Chapter 5
in volume 1, Figure 5.6). The two shares are not the same, because the
distributional figures concern households. Between households and the total
economy stand various institutions, including the company sector (which
retains profits), pension funds (which own shares), and the government
(which levies taxes and receives profit income). The dividends paid to
pension funds, for example, generate the income which is then paid to
65
pensioners, in whose hands it is treated as deferred earnings, so that – in
these statistics – it does not appear as unearned income. It is nonetheless
interesting to examine the relation between factor shares and top incomes.
The separation of national and household income is one reason why
the decline of top capital incomes may have taken place even if the factor
share of capital has remained unchanged. This point is made forcefully for
France by Piketty (Chapter 3 in volume 1). Profits may be retained within
the company sector and rents may be accruing to owner-occupiers or public
authorities rather than to private landlords. (These are, of course, a
reminder of the incompleteness of the measure of income in the income tax
data.) On the other hand, in some other countries there is a correlation.
Roine and Waldenström plot for Sweden the changes in the capital share of
value added and the evolution of the top one percent income share. The
series are strongly correlated over the whole period, but with a clear
difference between the first and second half of the century. Between 1907
and 1950 the correlation is 0.94, while it drops to 0.55 between 1951 and
2000. This indicates that, at least during the first fifty years, even short
term fluctuations of top incomes follow the fluctuations of the capital share
of value added as a share of GDP. They also find a downward trend in the
capital share of value added over the first 80 years.
Global forces
While popular theories of income distribution concentrate on a closed
economy, top income shares are undoubtedly influenced by international
movements of capital and labour. The extent of mobility has differed over
66
time, and our observations span a wide variety of periods, including the
previous globalization of the nineteenth century and the protectionism of
the interwar years.
It would clearly be interesting to use the data contained in the two
volumes covering twenty two countries, with much of the data on a near-
annual basis, to explore the common economic influences on the evolution
of top shares and possible interdependencies. Important among the common
forces are the degree of integration of capital markets and the movements
in major commodity prices.
One line of approach is to contrast the time variation of different
income groups. A common feature to most of the chapters has been the
difference between the time paths of the very top groups and the paths
followed by those just below the top. The top 1 per cent, and certainly the
top 0.1 per cent, are different from the next 9 per cent (9.9 per cent). It is
indeed interesting to ask whether the top 0.1 per cent are more like their
counterparts in other countries than they are like the next 9.9 per cent in
their own country.
If we consider possible explanatory variables, then the most obvious
candidates are the rate of return, movements in commodity prices (to which
we have already made references), and, in recent years, the international
market for managers and for superstars.
In addition to global correlations, there are other cross-country
commonalities. Saez and Veall (Chapter 6 in volume 1) use the top income
share in the US as an explanatory variable in a regression explaining the top
income share in Canada. Leigh and van der Eng (Chapter 4) show the
67
correlation between the top income share in Indonesia and those in other
countries. They conclude that the correlation is highest with another
developing country – India – but note that the correlation with Argentina is
negative.
Progressive taxation
In the study of France that initiated this project, Piketty (2001 and 2003)
highlighted the role of progressive income taxation: “how can one account
for the fact that large fortunes never recovered from the 1914-45 shocks,
while smaller fortunes did recover perfectly well? The most natural and
plausible candidate for an explanation seems to be the creation and
development of the progressive income tax” (Chapter 3 in Volume 1, page
61). It should be stressed here that this conclusion refers to the impact on
the distribution of gross income: i. e. income before the deduction of
income tax. (See Table 4.2 on the UK in volume 1 for one of the few tables
that relate to the distribution of income after tax.)
Evidence about the impact of taxation is discussed in many of the
chapters. In the case of Sweden, Roine and Waldenström conclude that
“Progressive taxation hence seems to have been a major contributing factor
in explaining the evolution of Swedish top incomes in the postwar period.
However, given that much of the fall in top incomes happens before taxes
reach extreme levels and largely as a result of decreasing income from
wealth, an important effect of taxation in terms of top income shares has
been to prevent the accumulation of new fortunes.” (Chapter 7, page xxx).
In the case of Finland, Jäntti, Riihelä, Sullström and Tuomala conclude that
68
the decline in income tax progressivity since the mid 1990s is a central
factor explaining the increase of top income shares in Finland. In the case of
Switzerland, a country that has never having imposed very high rates of
taxation, Dell, Piketty and Saez (Chapter 11 in volume 1) conclude that the
observed stability of top shares is consistent with the explanation of trends
elsewhere in terms of tax effects.
Outside Europe, Moriguchi and Saez recall in the case of Japan “that
the enormous fortunes that generated the high top 1% income share in the
pre-Second World War period had been accumulated at the time when
progressive income tax hardly existed and capitalists could reinvest almost
all of their incomes for further capital accumulation” (Chapter 3, page xxx).
They go on to say that the fiscal environment faced by Japanese capitalists
after the Second World War was vastly different: the top statutory marginal
tax rate for individual income tax stayed at 60-75% from 1950 until the 1988
tax reform. Progressive taxation hindered the re-accumulation of large
wealth, resulting in more equal distribution of capital income. This is the
same mechanism that Piketty had earlier identified in France, and was
highlighted in the case of the US by Piketty and Saez (Chapter 5 in volume
1). Noting that “it is difficult to prove in a rigorous way that the dynamic
effects of progressive taxation on capital accumulation and pre-tax
inequality have the right quantitative magnitude and account for the
observed facts” (volume 1, page 157), they conclude that the interpretation
seems reasonable on a priori grounds.
On the other hand, there are different findings in some countries.
Saez and Veall devote a whole section of their study of Canada (Chapter 6 in
69
volume 1) to the role of taxation and the consequences of the drop in
marginal tax rates since the 1960s. They conclude that “the concentration
of the surge in the last decade and among only the very top income shares
suggests that tax changes in Canada cannot be the sole cause.” (Chapter 6
in volume 1, page 257). Their econometric analysis finds that “Canadian top
income changes are much more strongly associated with similar US changes
than with Canadian tax developments” (Chapter 6 in volume 1, page 257).
The econometric research of Leigh and van der Eng for Indonesia (Chapter
4) does not find conclusive evidence of a link with marginal tax rates. In
the chapter on Portugal, Alvaredo notes the top tax rate has been constant
(Figure 11.4) at a new lower rate for a long period, during which top shares
continued to rise. The same is true for the UK (Chapter 4 in volume 1),
where top shares rose steadily over the 20 years since the top rate of
income tax was reduced to 40 per cent. As noted by Saez, ”in contrast to
the United States ... the increase in top share has been relatively smooth
since 1979 with no break around the tax changes” (2004, page 33).
As these latter cases bring out, a key element in assessing the effect
of taxation concerns the timing of the impact. Is the current income share a
function of the current tax rate or of the past tax rates? The answer
depends on how we envisage the underlying behavioural model. The models
used by Saez (2004) to examine the relation between marginal tax rates and
reported incomes are based on current tax rates. In Chapter 10, Alvaredo
and Saez examine the response of business organisation to taxation using a
model that relates current incomes to current tax rates. On the other hand,
models of wealth accumulation typically treat the change in wealth as a
70
function of the current tax rate. In this case, the present top income shares
may reflect a weighted average of past tax rates. Piketty (2001 and 2003)
provides numerical simulations with a fixed saving rate model, which
indicate that substantial capital taxes are a serious obstacle to the recovery
of wealth holdings from negative shocks, and that the barriers would be
further raised if the reduction in the rate of return were to reduce the
propensity to save.
Summary
We have sketched in this section some of the major mechanisms influencing
the development of top income shares. Understanding the relative
importance of these different factors is important in the design of public
policy. Concern about the rise in top shares in a number of countries has
led to a range of proposals. Some countries have already announced
increases in top income tax rates; others are considering limits on
remuneration. These are being implemented at a time of recession, which
may too lead to a decline in top shares.
13.6 ENVOI
The subtitle of this volume – A Global Perspective – is an exaggeration.
Major countries are missing, such as Brazil and Russia; we have no evidence
for Africa and Latin America is represented only by Argentina. At the same
time, the 22 countries covered in the two volumes contain more than half
(54 per cent) of the world’s population. Our data cover much of the
twentieth century, including the Great Depression, the Golden Age, and the
71
Roaring Nineties. In some cases, the data reach back before the First World
War and into the nineteenth century. We hope that the data will provide a
rich source for future researchers.
REFERENCES
Atkinson, A. B. and Brandolini, A. (2006). “The panel-of-countries approach
to explaining income inequality” in S. L. Morgan, D. B. Grusky and G. S
Fields, editors, Mobility and Inequality. Stanford: Stanford University Press.
Atkinson, A. B. and Harrison, A. J. (1978). The Distribution of Personal
Wealth in Britain. Cambridge: Cambridge University Press.
Atkinson, A. B. and Leigh, A. (2007a). “The distribution of top incomes in five
Anglo-Saxon countries over the twentieth century”. Canberra: Australian
National University, mimeo.
Atkinson, A. B. and Leigh, A. (2008). “Top Incomes in New Zealand 1921-2005:
Understanding the Effects of Marginal Tax Rates, Migration Threat, and the
Macroeconomy’. Review of Income and Wealth, 54(2): 149-165
Atkinson, A. B. and Piketty, T. (2007). Top incomes over the twentieth
century: A contrast between Continental European and English-Speaking
countries. Oxford: Oxford University Press.
Autor, D. H., Katz, L. F. and Kearney, M. S. (2006). “The polarization of the
U.S. labour market”, American Economic Review, Papers and Proceedings,
vol 96: 189-194.
72
Bach, S., Corneo, G. and Steiner, V. (2008). “Effective taxation of top
incomes in German, 1992-2002”. DIW Discussion Paper 767, Berlin.
Brewer, M., Muriel, A., Phillips, D. and Sibieta, L. (2008). Poverty and
Inequality in the UK: 2008. London: Institute for Fiscal Studies.
Coles, S. (2001). An Introduction to Statistical Modeling of Extreme Values.
London: Springer Verlag.
Dell, F. (2008). L’Allemagne Inégale. Ph D thesis, Paris School of Economics.
Embrechts, P., Klüppelberg, C. and Mikosch, T. (1997). Modelling Extremal
Events. Berlin: Springer Verlag.
Feinstein, C. H. (1972). Statistical Tables of National Income, Expenditure and
Output of the U.K. 1855-1965. Cambridge: Cambridge University Press.
Frank, R. H. and Cook, P. J. (1995). The Winner Take-All Society. New York:
Free Press.
Goode, R. (1964). The Individual Income Tax. Washington, D. C.: Brookings
Institution.
Gordon, R. and J. Slemrod (2000). “Are Real Responses to Taxes Simply
Income Shifting Between Corporate and Personal Tax Bases?,” in Slemrod J.
(ed.) Does Atlas Shrug? The Economic Consequences of Taxing the Rich,
Cambridge: Cambridge University Press.
Goldin, C. and Margo, R. (1992). “The Great Compression: The wage structure
in the United States at mid-century”, Quarterly Journal of Economics, 107: 1-
34.
Guilera, J. (2008). “Top Income Shares in Portugal over the Twentieth
Century”, Documents de Treball de la Facultat de Ciències Econòmiques I
Empresarials, E08/195, Universitat de Barcelona.
73
Gustafsson, B. and Jansson, B. “Top incomes in Sweden during three-
quarters of a century: A micro-data approach”, IZA Discussion Paper 2672,
IZA Bonn.
Katz, L. F. and Autor, D. H. (1999). “Changes in the wage structure and
earnings inequality” in O. Ashenfelter and D. Card, editors, Handbook of
Labor Economics, volume 3A. Amsterdam: North-Holland.
Kuznets, S. (1953). Shares of Upper Income Groups in Income and Savings.
New York: National Bureau of Economic Research.
Kuznets, S. (1955). “Economic Growth and Income Inequality”, American
Economic Review, 45: 1-28.
Landais, C. (2007). “Les hauts revenus en France 1998-2006: Une explosion
des inégalités?” [Top incomes in France 1998-2006: An explosion of
inequalities?] Paris School of Economics Working Paper.
Lazear, E. P. and Rosen, S. (1981). “Rank-order tournaments as optimum labor
contracts”, Journal of Political Economy, vol 89: 841-864.
Leigh, A. (2007). “How Closely do Top Income Shares Track Other Measures of
Inequality?”, Economic Journal, 117: F619-F633.
Leigh, A. (2009). “Top Incomes”, in W. Salverda, B. Nolan, and T. Smeeding
(editors) The Oxford Handbook of Economic Inequality. Oxford: Oxford
University Press .
Lydall, H. F. (1959). “The Distribution of Employment Incomes”,
Econometrica, vol 27: 110-115.
Marshall, A. (1920). Principles of Economics, Eighth edition. London:
Macmillan.
74
Meade, J. E. (1964). Efficiency, Equality and the Ownership of Property.
London: Allen and Unwin.
Merz, J., Hirschel, D. and Zwick, M. (2005). “Struktur und Verteilung hoher
Einkommen – Mikroanalysen auf der Basis der Einkommensteuerstatistik”,
Beitrag zum zweiten Armuts- und Reichtumsbericht 2004 der Bundesregierung.
Morgenstern, O. (1963). On the Accuracy of Economic Observations, second
edition. Princeton: Princeton University Press.
Piketty, T. (2001). Les hauts revenus en France au 20ème siècle. Paris: Grasset.
Piketty, T. (2003). “Income Inequality in France, 1901-1998”, Journal of
Political Economy, vol 111: 1004-1042.
Piketty, T., G. Postel-Vinay and J.L. Rosenthal, “Wealth Concentration in a
Developing Economy: Paris and France, 1807-1994”, American Economic
Review 2006
Piketty, T. and Saez, E. (2006). “The evolution of top incomes: A historical
and international perspective”, American Economic Review, Papers and
Proceedings, 96: 200-5.
Roine, J., Vlachos, and Waldenström, D. (2008). “The long-run determinants
of inequality: What can we learn from top income data?” Discussion paper.
Rosen, S. (1981). "The Economics of Superstars", American Economic Review,
vol 71: 845-858.
Saez, E. (2004). “Reported Incomes and Marginal Tax Rates, 1960-2000:
Evidence and Policy Implications”, in J. Poterba (editor) Tax Policy and the
Economy, 18, Cambridge: MIT Press.
75
Saez, E., Slemrod, J., and Giertz, S. (2009). “The Elasticity of Taxable
Income with Respect to Marginal Tax Rates: A Critical Review”, working
paper February 2009, in preparation for the Journal of Economic Literature.
Scheve, K. and Stasavage, D. (2009). “Institutions, partisanship, and inequality
in the long run”, World Politics, vol 61:
Simon, H. (1957). “The Compensation of Executives”, Sociometry, vol 20: 32-
35.
Stiglitz, J. E. (1969). “Distribution of income and wealth among individuals”,
Econometrica, vol 37: 382-397.
Tinbergen, J. (1975). Income distribution: Analysis and policies.
Amsterdam: North-Holland.
Titmuss, R. M. (1962). Income Distribution and Social Change. London: Allen
and Unwin.
Vaughan, R. N. (1993). “On the microfoundations of wealth distribution
functions”, in E. N. Wolff (editor) Research on Economic Inequality, Volume 4:
Studies in the Distribution of Household Wealth. Greenwich: JAI Press.
Figure 13.1 Coverage of countries and years
1875 1885 1895 1905 1915 1925 1935 1945 1955 1965 1975 1985 1995 2005
Norway
Japan
Prussia/Germany
Sweden
France
UK
US
Netherlands
Canada
Finland
Indonesia
Australia
New Zealand
India
Ireland
Argentina
Spain
Switzerland
Portugal
Singapore
Italy
China
Figure 13.2 Effect of capital gains on share of top 1 per cent
0
5
10
15
20
25
1949
1951
1953
1955
1957
1959
1961
1963
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
Shar
e in
tota
l inc
ome
of to
p 1
per c
ent
US US inc CGs
CA CA inc CGs
SWE SWE inc CGs
FI FI inc CGs
ESP ESP inc CGs
Table 13.1 Comparative top income sharesAround 1949 Around 2005
share of top 1%
share of top 0.1%
β coefficient
share of top 1%
share of top 0.1%
β coefficient
Indonesia 19.87 7.03 2.22 1.34 2.94Argentina 19.34 7.87 2.56 16.75 7.02 2.65Ireland 12.92 4.00 1.96 10.30 2.00Netherlands 12.05 3.80 2.00 5.38 1.08 1.43India 12.00 5.24 2.78 8.95 3.64 2.56Germany 11.60 3.90 2.11 11.10 4.40 2.49United Kingdom 11.47 3.45 1.92 14.25 5.19 2.28Australia 11.26 3.31 1.88 8.79 2.68 1.94United States 10.95 3.34 1.94 17.42 7.70 2.82Canada 10.69 2.91 1.77 13.56 5.23 2.42Singapore 10.38 3.24 1.98 13.28 4.29 2.04New Zealand 9.98 2.42 1.63 8.76 2.51 1.84Switzerland 9.88 3.23 2.06 7.76 2.67 2.16France 9.01 2.61 1.86 8.20 2.19 1.74Norway 8.88 2.74 1.96 11.82 5.59 3.08Japan 7.89 1.82 1.57 9.20 2.40 1.71Finland 7.71 1.63 7.08 2.65 2.34Sweden 7.64 1.96 1.69 6.28 1.91 1.93Spain 1.99 8.79 2.62 1.90Portugal 3.57 1.94 9.13 2.26 1.65Italy 9.03 2.55 1.82China 5.87 1.20 1.45
Notes:(1) 1939 for Indonesia, 1943 for Ireland,1950 for Germany and the Netherlands, 1954 for Spain
(2) 1995 for Switzerland, 1998 for Germany, 1999 for Netherlands, 1999-2000 for India2000 for Canada and Ireland, 2002 for Australia, 2003 for Indonesia and Portugal2004 for Argentina, Italy, Norway and Sweden
(3) β coefficients are calculated using share of top 0.1% in top 1% (see Tables 13A.23 and 13A.24), with the following exceptions:(i) β coefficient for Finland in 1949 calculated using share of top 1% in top 5%(ii) β coefficient for Spain in 1949 calculated using share of top 0.01% in top 0.05%(iii) β coefficient for Portugal in 1949 calculated using share of top 0.01% in top 0.1%(iv) β coefficient for Ireland in 2000 calculated using share of top 0.5% in top 1%(v) β coefficient for Indonesia in 2003 calculated using share of top 0.01% in top 0.1%
Table 13.1B: Pareto-Lorenz α coefficients vs. inverted-Pareto-Lorenz β coefficients
α β = α/(α-1) β α = β/(β-
1)
1.10 11.00 1.50 3.001.30 4.33 1.60 2.671.50 3.00 1.70 2.431.70 2.43 1.80 2.251.90 2.11 1.90 2.112.00 2.00 2.00 2.002.10 1.91 2.10 1.912.30 1.77 2.20 1.832.50 1.67 2.30 1.773.00 1.50 2.40 1.714.00 1.33 2.50 1.675.00 1.25 3.00 1.5010.00 1.11 3.50 1.40
Notes: (i) The "α" coefficient is the standard Pareto-Lorenz coefficient commonly usedin power-law distribution formulas: 1-F(y) = (A/y)α and f(y) = αAα/y1+α (A>0, α>1, f(y) = density function, F(y) = distribution function, 1-F(y) = proportion of population with income above y). A higher coefficient α means a faster convergence of the density towards zero, i.e. a less fat upper tail. (ii) The "β" coefficient is defined as the ratio y*(y)/y, i.e. the ratio between the average income y*(y) of individuals with income above threshold y and the threshold y. The characteristic property of power laws is that this ratio is a constant, i.e. does not dependon the threshold y. Simple computations show that β = y*(y)/y = α/(α-1), and converselyα = β/(β-1).
Table 13.2 Summary of changes in shares of top 1 per cent and 0.1 per cent between 1949 and 2005
Country Share of top 1 per cent Share of top 0.1 per cent
FranceNo change. Rose 1 point between 1998 and 2005.
Fell 1 point between 1949 and early 1980s. Rose 0.4 point between 1998 andf 2005.
UK Fell 6; rose 7½ points. Fell 2; rose 3 points.US Fell 3; rose 10 points. Fell 1; rose 6 points.Canada Fell 3; rose 6 points (up to 2000). Fell 1; rose 3½ points (up to 2000).Australia Fell 7; rose 4 points. Fell 2; rose 1½ points.New`Zealand Fell 3; rose 4 points. Fell 1; rose 1½ points.Germany No sustained change. No sustained change.Netherlands Fell 6½ points (up to 1999). Fell 3 points (up to 1999).Switzerland No sustained change. No sustained change.India Fell 7½; rose 4½ points (up to 1999). Fell 4; rose 2½ points (up to 1999).Japan No sustained change up to 1999; rose 1½
points between 1999 and 2005.No sustained change up to 1999; rose ¾ point between 1999 and 2005.
SingaporeNo sustained change from 1960 to 1998; rose 2 points between 1998 and 2005.
No sustained change from 1960 to 1990s; rose 2 points between 1990s and 2005.
Argentina Fell 12; rose 4 points. Fell 5½; rose 3 points.Sweden Fell 3½; rose 2 points. Fell 1¼; rose 1¼ points.Finland Rose 2 points up to early 1960s; fell 6 points;
rose 3½ points.Norway Fell 4½; rose 8 points. Fell 1¾; rose 4½ points.
Notes(1) "No change" means change less than 2 percentage points for top 1 per cent; less than 0.65 percentage point for top 0.1 per cent.
(2) Data coverage incomplete for part of the period for Argentina
Figure 13.3. Inverted-Pareto-Lorenz β coefficients, 1949-2005: "flat" countries
1.0
1.5
2.0
2.5
3.0
3.5
194919
5119
5319
5519
5719
5919
6119
6319
6519
6719
6919
7119
7319
7519
7719
7919
8119
8319
8519
8719
8919
9119
9319
9519
9719
9920
0120
0320
05
Pare
to-L
oren
z co
effic
ient
FR
DE
NL
CH
JA
SI
Figure 13.4. Inverted-Pareto-Lorenz β coefficients, 1949-2005: U-shaped countries
1.0
1.5
2.0
2.5
3.0
3.5
194919
5119
5319
5519
5719
5919
6119
6319
6519
6719
6919
7119
7319
7519
7719
7919
8119
8319
8519
8719
8919
9119
9319
9519
9719
9920
0120
0320
05
Pare
to-L
oren
z co
effic
ient
UK
US
CA
AUS
NZ
IN
ARG
SE
NO
Table 13.3 Summary of changes in shares of top 1 per cent and 0.1 per cent before 1949
Country Share of top 1 per cent Share of top 0.1 per cent
France1928-31: lose 2 points 1928-31: lose a fifthWW2: lose 4 points WW2: halved1949 = half of 1914 1949 = a third of 1919
UK- WW1: lose a fifth- 1928-31: lose a fifth- WW2: lose 30 per cent
1949 = half of 1914 1949 = 40 per cent of 1919Pre-WW1: no obvious trend
US WW1: lose 3 points WW1: lose a third1928-31: lose 4 points 1928-31: lose a thirdWW2: lose 3 points WW2: lose a third1949 = 70 per cent of 1919 1949 = half of 1919
Canada1928-31: gain 1 point 1928-31: no changeWW2: lose 6 points WW2: halved1949 = ¾ of 1920 1949 = half of 1920
Australia1928-31: lose 2½ points 1928-31: lose a quarterWW2: lose 1 point WW2: lose a quarter1949 same as 1921 1949 = 85 per cent of 1921
New`Zealand1928-30: lose 1 point 1928-30: lose a fifthWW2: lose 2 points WW2: lose a quarter1949 = ⅔ of 1921 1949 = half of 1921
Germany1928-32: no change 1928-32: no change1933-38: gain 5 points 1933-38: gain 3 points1950 = ⅔ of 1938 1950 = half of 1938Prussia: 1914 unchanged relative to 1881 Prussia: 1914 unchanged relative to 1881(Germany 1925 = 60% of Prussia 1914) (Germany 1925 = half of Prussia 1914)
NetherlandsWW1: gain 3 points WW1: gain a quarter1928-32: lose 4 points 1928-32: lose a thirdWW2: lose 5 points WW2: lose a third1950 = 60 per cent of 1914 1950 = 45 per cent of 1914
SwitzerlandWW2: lose 1 point WW2: lose a fifth
1949 is unchanged relative to 1933 1949 is unchanged relative to 1933
Ireland28-32: gain 40 per centWW2: lose a fifth1949 same as 1922
India28-31: gain 2 points 28-31: gain a fifth WW2: lose 5 points WW2: lose a quarter1949 is unchanged relative to 1922 1949 is unchanged relative to 1922
JapanWW1: lose 3 points WW1: lose a tenth28-31: lose 1 point 28-31: lose a tenthWW2: lose 9 points WW2: lose two-thirds1949 = 40 per cent of 1914 1949 = quarter of 19141914 is unchanged relative to 1886 1914 is unchanged relative to 1886
Indonesia28-32: gain 5 points 28-32: gain 15 per cent1939 = 8 points higher than 1921 1939 = quarter higher than 1921
ArgentinaWW2: gain of 2 points WW2: gain of fifth1949 is unchanged relative to 1932 1949 is unchanged relative to 1932
Sweden1949 is a third of 1912 1949 is a fifth of 19121912 = ¾ of 1903 1912 unchanged relative to 1903
Finland28-30: no changeWW2: loss of 5 points1949 = half 1920
NorwayWW2: lose 4 points WW2: lose 40 per cent1949 = ¾ of 19131913 = ⅔ of 1875
Spain1949 = 60 per cent of 1933
Portugal1949 = ⅔ of 1936
Notes(1) WW1 denotes the First World War; WW2 denotes the Second World War(2) "No change" means change less than 2 percentage points for top 1 per cent; less than 0.65 percentage point for top 0.1 per cent.
(2) Data coverage incomplete for part of the period for Argentina
Table 13.4 Summary of changes in shares of top "next 4 per cent" and "second vintile"
Country "Next 4 per cent" "Second vintile" Text comments
France1919 14.3 1919 8.41949 12.7 1949 10.52005 13.0 2005 11.0
UK1919 11.9 1919 7.21949 11.9 1949 8.91978 11.4 1978 10.72005 14.5 2005 11.2
US1919 13.5 1919 10.21949 12.5 1949 10.32005 15.2 2005 11.8
Canada1920 18.21949 14.7 1949 12.82000 15.4 2000 13.3
Australia1921 7.81949 12.4 1949 9.12002 11.2 2002 10.4
New`Zealand1921 14.11949 12.3 1949 9.22005 12.7 2005 10.8
Germany1950 13.3 1950 9.51998 13.1 1998 11.2
"The secular decline of the top decile income share is almost entirely due to very high incomes" (Vol 1, 48).
"The highlights the 'localised nature of redistribution'" (Vol 1, 96).
After 1958, "the downward trend continued for the next 4% but not for the second vintile" (Vol 1, 320).
The "upturn during the last two decades is concentrated in the top percentile" (Vol 1, 232).
After 1953, "the share of the [second] vintile was not much reduced" (Vol 1, 343).
The next 4% and the second vintile "account for a relatively small fraction of the total fluctuation of the top decile income share" (Vol 1, 146).
"The bottom part of the top decile does not exhibit the same stability as the upper part. … From the early 1960s … the share of the bottom 9% of the top decile has been constantly growing" (Vol 1, 377).
Netherlands1919 15.7 1919 10.11950 14.1 1950 10.61999 11.7 1999 11.0
Switzerland1949 12.3 1949 10.11995 11.5 1995 9.9
Ireland(next 9%) 1943 30.3 -
2000 25.8 -
China1986 7.2 1986 7.62003 11.9 2003 10.2
Japan1919 9.6 -1949 13.8 -2005 16.1 -
Singapore1974 12.3 1974 7.92005 14.6 2005 9.5
Sweden1919 14.9 1919 10.71949 12.3 1949 10.52005 11.1 2005 9.6
Finland1920 18.3 -
"The two bottom groups [the next 4% and the second vintile] are remarkably stable over the period" (Vol 1, 488).
"Most of the inter-war decline of the top 10% is restricted to the top 1%, while its post-war decline is broader and covers the upper vintile as a whole" (Vol 1, 444).
"a much sharper rise [from 1990 to 2000] the higher one goes up the distribution" (Vol 1, 515).
"the rise in income inequality was so much concentrated within top incomes in both countries [China and India]" (xxx).
"the income de-concentration phenomenon that took place during the Second World War was limited to within the top 1% …[From 1992 to 2005 there has been] a sharp increase [in the share of the next 4%]" (xxx).
"Looking first at the decline over the first eighty years of the century, we see that virtually all of the fall in the top decile income share is due to a decrease in the very top of the distribution. The income share for the lower half of the top decile (P90–95) has been remarkably stable" (xxx).
"Over a thirty year period there was broad stability of the very top income shares. Ar the same time there was some change lower down the distribution" (xxx).
"Compared with top one per cent group, the income shares of
1949 13.0 -1992 12.1 -1965 10.7 1965 9.82004 9.5 2004 8.7
Norway1913 12.4 1913 9.31949 13.2 1949 11.92005 11.3 2005 9.4
Spain1981 13.6 1981 11.52005 13.4 2005 11.0
Portugal1976 13.2 1976 10.62003 15.6 2003 11.7
Italy1974 12.4 1974 10.62004 12.3 2004 10.3
"Whereas the share of the top 1 per cent rose by some 7 percentage points between 1991 and 2004, the share of the next 4 per cent increased by only about 2 percentage points, and there was virtually no rise in the share of those in the [second vintile]" (xxx).
"in Portugal, all groups within the top decile display important increases" (xxx).
"the increase in income concentration which took place in Italy since the mid 1980s has been a phenomenon happening within the top 5% of the distribution" (xxx).
"the increase in income concentration which took place in Spain since 1981 has been a phenomenon concentrated within the top 1% of the distribution" (xxx).
percentile groups within the rest of the 10 per cent has risen relatively modestly over the last ten years".(xxx)
Figure 13.5. Top 1% income shares, 1900-2005: L-shaped countries
0.0
5.0
10.0
15.0
20.0
25.0
30.0
1900
1904
1908
1912
1916
1920
1924
1928
1932
1936
1940
1944
1948
1952
1956
1960
1964
1968
1972
1976
1980
1984
1988
1992
1996
2000
2004
Pare
to-L
oren
z co
effic
ient
FR
DE
NL
CH
IR
JA
SI
Figure 13.6. Inverted-Pareto-Lorenz β coefficients, 1900-2005: L-shaped countries
1.0
1.5
2.0
2.5
3.0
3.5
4.0
1900
1904
1908
1912
1916
1920
1924
1928
1932
1936
1940
1944
1948
1952
1956
1960
1964
1968
1972
1976
1980
1984
1988
1992
1996
2000
2004
Pare
to-L
oren
z co
effic
ient
FR
DE
NL
CH
IR
JA
SI
Figure 13.7. Top 1% income shares, 1900-2005: U-shaped countries
0.0
5.0
10.0
15.0
20.0
25.0
30.0
1900
1904
1908
1912
1916
1920
1924
1928
1932
1936
1940
1944
1948
1952
1956
1960
1964
1968
1972
1976
1980
1984
1988
1992
1996
2000
2004
Shar
e of
top
1% in
tota
l inc
ome
UK
US
CA
AUS
NZ
IN
ARG
SE
NO
Figure 13.8. Inverted-Pareto-Lorenz β coefficients, 1900-2005: U-shaped countries
1.0
1.5
2.0
2.5
3.0
3.5
4.0
1900
1904
1908
1912
1916
1920
1924
1928
1932
1936
1940
1944
1948
1952
1956
1960
1964
1968
1972
1976
1980
1984
1988
1992
1996
2000
2004
Pare
to-L
oren
z co
effic
ient
UK
US
CA
AUS
NZ
IN
ARG
SE
NO